Merge pull request #57 from BobbyL2k/feat/llama-cpp-proxy

feat: Proxy llama.cpp API endpoints via `/llama-cpp/{name}/`

.dockerignore

@@ -0,0 +1,45 @@
+# Git and version control
+.git/
+.gitignore
+
+# Documentation
+*.md
+docs/
+
+# Development files
+.vscode/
+.idea/
+
+# Build artifacts
+webui/node_modules/
+webui/dist/
+webui/.next/
+*.log
+*.tmp
+
+# Data directories
+data/
+models/
+logs/
+
+# Test files
+*_test.go
+**/*_test.go
+
+# CI/CD
+.github/
+
+# Local configuration
+llamactl.yaml
+config.yaml
+.env
+.env.local
+
+# OS files
+.DS_Store
+Thumbs.db
+
+# Backup files
+*.bak
+*.backup
+*~

.github/workflows/docs.yml

@@ -0,0 +1,103 @@
+name: Build and Deploy Documentation
+
+on:
+  push:
+    branches: [ main ]
+    tags: [ 'v*' ]
+  pull_request:
+    branches: [ main ]
+    paths:
+      - 'docs/**'
+      - 'mkdocs.yml'
+      - 'docs-requirements.txt'
+
+permissions:
+  contents: write
+  pages: write
+  id-token: write
+
+concurrency:
+  group: "pages"
+  cancel-in-progress: false
+
+jobs:
+  build:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
+
+      - name: Setup Python
+        uses: actions/setup-python@v4
+        with:
+          python-version: '3.11'
+
+      - name: Install dependencies
+        run: |
+          pip install -r docs-requirements.txt
+
+      - name: Build documentation
+        run: |
+          mkdocs build --strict
+
+  deploy-dev:
+    runs-on: ubuntu-latest
+    if: github.ref == 'refs/heads/main' && github.event_name == 'push'
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
+
+      - name: Setup Python
+        uses: actions/setup-python@v4
+        with:
+          python-version: '3.11'
+
+      - name: Install dependencies
+        run: |
+          pip install -r docs-requirements.txt
+
+      - name: Configure Git
+        run: |
+          git config --global user.name "${{ github.actor }}"
+          git config --global user.email "${{ github.actor }}@users.noreply.github.com"
+
+      - name: Deploy development version
+        run: |
+          mike deploy --push --update-aliases dev latest
+          # Set dev as default if no default exists
+          if ! mike list | grep -q "default"; then
+            mike set-default --push dev
+          fi
+
+  deploy-release:
+    runs-on: ubuntu-latest
+    if: startsWith(github.ref, 'refs/tags/v')
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
+
+      - name: Setup Python
+        uses: actions/setup-python@v4
+        with:
+          python-version: '3.11'
+
+      - name: Install dependencies
+        run: |
+          pip install -r docs-requirements.txt
+
+      - name: Configure Git
+        run: |
+          git config --global user.name "${{ github.actor }}"
+          git config --global user.email "${{ github.actor }}@users.noreply.github.com"
+
+      - name: Deploy release version
+        run: |
+          VERSION=${GITHUB_REF#refs/tags/}
+          mike deploy --push --update-aliases $VERSION stable
+          mike set-default --push stable

.gitignore

@@ -32,4 +32,6 @@ go.work.sum
 # .vscode/
 
 node_modules/
-dist/
+dist/
+
+__pycache__/

.vscode/launch.json

@@ -0,0 +1,19 @@
+{
+    // Use IntelliSense to learn about possible attributes.
+    // Hover to view descriptions of existing attributes.
+    // For more information, visit: https://go.microsoft.com/fwlink/?linkid=830387
+    "version": "0.2.0",
+    "configurations": [
+        {
+            "name": "Launch Server",
+            "type": "go",
+            "request": "launch",
+            "mode": "auto",
+            "program": "${workspaceFolder}/cmd/server/main.go",
+            "env": {
+                "GO_ENV": "development",
+                "LLAMACTL_REQUIRE_MANAGEMENT_AUTH": "false"
+            },
+        }
+    ]
+}

CONTRIBUTING.md

@@ -129,6 +129,50 @@ Use this format for pull request titles:
 - Use meaningful component and variable names
 - Prefer functional components over class components
 
+## Documentation Development
+
+This project uses MkDocs for documentation. When working on documentation:
+
+### Setup Documentation Environment
+
+```bash
+# Install documentation dependencies
+pip install -r docs-requirements.txt
+```
+
+### Development Workflow
+
+```bash
+# Serve documentation locally for development
+mkdocs serve
+```
+The documentation will be available at http://localhost:8000
+
+```bash
+# Build static documentation site
+mkdocs build
+```
+The built site will be in the `site/` directory.
+
+### Documentation Structure
+
+- `docs/` - Documentation content (Markdown files)
+- `mkdocs.yml` - MkDocs configuration
+- `docs-requirements.txt` - Python dependencies for documentation
+
+### Adding New Documentation
+
+When adding new documentation:
+
+1. Create Markdown files in the appropriate `docs/` subdirectory
+2. Update the navigation in `mkdocs.yml`
+3. Test locally with `mkdocs serve`
+4. Submit a pull request
+
+### Documentation Deployment
+
+Documentation is automatically built and deployed to GitHub Pages when changes are pushed to the main branch.
+
 ## Getting Help
 
 - Check existing [issues](https://github.com/lordmathis/llamactl/issues)

README.md

@@ -2,30 +2,39 @@
 
 ![Build and Release](https://github.com/lordmathis/llamactl/actions/workflows/release.yaml/badge.svg) ![Go Tests](https://github.com/lordmathis/llamactl/actions/workflows/go_test.yaml/badge.svg) ![WebUI Tests](https://github.com/lordmathis/llamactl/actions/workflows/webui_test.yaml/badge.svg)
 
-**Management server and proxy for multiple llama.cpp instances with OpenAI-compatible API routing.**
+**Unified management and routing for llama.cpp, MLX and vLLM models with web dashboard.**
 
-## Why llamactl?
+## Features
 
-🚀 **Multiple Model Serving**: Run different models simultaneously (7B for speed, 70B for quality)  
-🔗 **OpenAI API Compatible**: Drop-in replacement - route requests by model name  
-🌐 **Web Dashboard**: Modern React UI for visual management (unlike CLI-only tools)  
-🔐 **API Key Authentication**: Separate keys for management vs inference access  
-📊 **Instance Monitoring**: Health checks, auto-restart, log management  
-⏳ **Idle Timeout Management**: Automatically stop idle instances after a configurable period  
-💡 **On-Demand Instance Start**: Automatically launch instances upon receiving OpenAI-compatible API requests  
-💾 **State Persistence**: Ensure instances remain intact across server restarts  
+### 🚀 Easy Model Management
+- **Multiple Model Serving**: Run different models simultaneously (7B for speed, 70B for quality)
+- **On-Demand Instance Start**: Automatically launch instances upon receiving API requests
+- **State Persistence**: Ensure instances remain intact across server restarts
 
-![Dashboard Screenshot](docs/images/screenshot.png)
+### 🔗 Universal Compatibility
+- **OpenAI API Compatible**: Drop-in replacement - route requests by instance name
+- **Multi-Backend Support**: Native support for llama.cpp, MLX (Apple Silicon optimized), and vLLM
+- **Docker Support**: Run backends in containers
 
-**Choose llamactl if**: You need authentication, health monitoring, auto-restart, and centralized management of multiple llama-server instances  
-**Choose Ollama if**: You want the simplest setup with strong community ecosystem and third-party integrations  
-**Choose LM Studio if**: You prefer a polished desktop GUI experience with easy model management
+### 🌐 User-Friendly Interface
+- **Web Dashboard**: Modern React UI for visual management (unlike CLI-only tools)
+- **API Key Authentication**: Separate keys for management vs inference access
+
+### ⚡ Smart Operations
+- **Instance Monitoring**: Health checks, auto-restart, log management
+- **Smart Resource Management**: Idle timeout, LRU eviction, and configurable instance limits
+- **Environment Variables**: Set custom environment variables per instance for advanced configuration  
+
+![Dashboard Screenshot](docs/images/dashboard.png)
 
 ## Quick Start
 
 ```bash
-# 1. Install llama-server (one-time setup)
-# See: https://github.com/ggml-org/llama.cpp#quick-start
+# 1. Install backend (one-time setup)
+# For llama.cpp: https://github.com/ggml-org/llama.cpp#quick-start
+# For MLX on macOS: pip install mlx-lm
+# For vLLM: pip install vllm
+# Or use Docker - no local installation required
 
 # 2. Download and run llamactl
 LATEST_VERSION=$(curl -s https://api.github.com/repos/lordmathis/llamactl/releases/latest | grep '"tag_name":' | sed -E 's/.*"([^"]+)".*/\1/')
@@ -42,15 +51,27 @@ llamactl
 ### Create and manage instances via web dashboard:
 1. Open http://localhost:8080
 2. Click "Create Instance"
-3. Set model path and GPU layers
-4. Start or stop the instance
+3. Choose backend type (llama.cpp, MLX, or vLLM)
+4. Set model path and backend-specific options
+5. Configure environment variables if needed (optional)
+6. Start or stop the instance
 
 ### Or use the REST API:
 ```bash
-# Create instance
+# Create llama.cpp instance
 curl -X POST localhost:8080/api/v1/instances/my-7b-model \
   -H "Authorization: Bearer your-key" \
-  -d '{"model": "/path/to/model.gguf", "gpu_layers": 32}'
+  -d '{"backend_type": "llama_cpp", "backend_options": {"model": "/path/to/model.gguf", "gpu_layers": 32}}'
+
+# Create MLX instance (macOS)
+curl -X POST localhost:8080/api/v1/instances/my-mlx-model \
+  -H "Authorization: Bearer your-key" \
+  -d '{"backend_type": "mlx_lm", "backend_options": {"model": "mlx-community/Mistral-7B-Instruct-v0.3-4bit"}}'
+
+# Create vLLM instance with environment variables
+curl -X POST localhost:8080/api/v1/instances/my-vllm-model \
+  -H "Authorization: Bearer your-key" \
+  -d '{"backend_type": "vllm", "backend_options": {"model": "microsoft/DialoGPT-medium", "tensor_parallel_size": 2}, "environment": {"CUDA_VISIBLE_DEVICES": "0,1", "NCCL_DEBUG": "INFO"}}'
 
 # Use with OpenAI SDK
 curl -X POST localhost:8080/v1/chat/completions \
@@ -74,7 +95,30 @@ sudo mv llamactl /usr/local/bin/
 # Windows - Download from releases page
 ```
 
-### Option 2: Build from Source
+### Option 2: Docker (No local backend installation required)
+
+```bash
+# Clone repository and build Docker images
+git clone https://github.com/lordmathis/llamactl.git
+cd llamactl
+mkdir -p data/llamacpp data/vllm models
+
+# Build and start llamactl with llama.cpp CUDA backend
+docker-compose -f docker/docker-compose.yml up llamactl-llamacpp -d
+
+# Build and start llamactl with vLLM CUDA backend
+docker-compose -f docker/docker-compose.yml up llamactl-vllm -d
+
+# Build from source using multi-stage build
+docker build -f docker/Dockerfile.source -t llamactl:source .
+```
+
+**Features:** CUDA support, automatic latest release installation, no backend dependencies.
+**Note:** Dockerfiles are configured for CUDA. Adapt base images for other platforms (CPU, ROCm, etc.).
+
+For detailed Docker setup and configuration, see the [Installation Guide](docs/getting-started/installation.md).
+
+### Option 3: Build from Source
 Requires Go 1.24+ and Node.js 22+
 ```bash
 git clone https://github.com/lordmathis/llamactl.git
@@ -85,16 +129,65 @@ go build -o llamactl ./cmd/server
 
 ## Prerequisites
 
+### Backend Dependencies
+
+**For llama.cpp backend:**
 You need `llama-server` from [llama.cpp](https://github.com/ggml-org/llama.cpp) installed:
 
 ```bash
-# Quick install methods:
 # Homebrew (macOS)
 brew install llama.cpp
 
 # Or build from source - see llama.cpp docs
+# Or use Docker - no local installation required
 ```
 
+**For MLX backend (macOS only):**
+You need MLX-LM installed:
+
+```bash
+# Install via pip (requires Python 3.8+)
+pip install mlx-lm
+
+# Or in a virtual environment (recommended)
+python -m venv mlx-env
+source mlx-env/bin/activate
+pip install mlx-lm
+```
+
+**For vLLM backend:**
+You need vLLM installed:
+
+```bash
+# Install via pip (requires Python 3.8+, GPU required)
+pip install vllm
+
+# Or in a virtual environment (recommended)
+python -m venv vllm-env
+source vllm-env/bin/activate
+pip install vllm
+
+# Or use Docker - no local installation required
+```
+
+## Backend Docker Support
+
+llamactl can run backends in Docker containers:
+
+```yaml
+backends:
+  llama-cpp:
+    docker:
+      enabled: true
+  vllm:
+    docker:
+      enabled: true
+```
+
+**Requirements:** Docker installed and running. For GPU support: nvidia-docker2 (Linux) or Docker Desktop with GPU support.
+
+For detailed Docker configuration options, see the [Configuration Guide](docs/getting-started/configuration.md).
+
 ## Configuration
 
 llamactl works out of the box with sensible defaults.
@@ -104,8 +197,35 @@ server:
   host: "0.0.0.0"                # Server host to bind to
   port: 8080                     # Server port to bind to
   allowed_origins: ["*"]         # Allowed CORS origins (default: all)
+  allowed_headers: ["*"]         # Allowed CORS headers (default: all)
   enable_swagger: false          # Enable Swagger UI for API docs
 
+backends:
+  llama-cpp:
+    command: "llama-server"
+    args: []
+    environment: {}               # Environment variables for the backend process
+    docker:
+      enabled: false
+      image: "ghcr.io/ggml-org/llama.cpp:server"
+      args: ["run", "--rm", "--network", "host", "--gpus", "all"]
+      environment: {}             # Environment variables for the container
+
+  vllm:
+    command: "vllm"
+    args: ["serve"]
+    environment: {}               # Environment variables for the backend process
+    docker:
+      enabled: false
+      image: "vllm/vllm-openai:latest"
+      args: ["run", "--rm", "--network", "host", "--gpus", "all", "--shm-size", "1g"]
+      environment: {}             # Environment variables for the container
+
+  mlx:
+    command: "mlx_lm.server"
+    args: []
+    environment: {}               # Environment variables for the backend process
+
 instances:
   port_range: [8000, 9000]       # Port range for instances
   data_dir: ~/.local/share/llamactl         # Data directory (platform-specific, see below)
@@ -113,7 +233,8 @@ instances:
   logs_dir: ~/.local/share/llamactl/logs    # Logs directory
   auto_create_dirs: true         # Auto-create data/config/logs dirs if missing
   max_instances: -1              # Max instances (-1 = unlimited)
-  llama_executable: llama-server # Path to llama-server executable
+  max_running_instances: -1      # Max running instances (-1 = unlimited)
+  enable_lru_eviction: true      # Enable LRU eviction for idle instances
   default_auto_restart: true     # Auto-restart new instances by default
   default_max_restarts: 3        # Max restarts for new instances
   default_restart_delay: 5       # Restart delay (seconds) for new instances
@@ -121,7 +242,6 @@ instances:
   on_demand_start_timeout: 120   # Default on-demand start timeout in seconds
   timeout_check_interval: 5      # Idle instance timeout check in minutes
 
-
 auth:
   require_inference_auth: true   # Require auth for inference endpoints
   inference_keys: []             # Keys for inference endpoints
@@ -129,103 +249,7 @@ auth:
   management_keys: []            # Keys for management endpoints
 ```
 
-<details><summary><strong>Full Configuration Guide</strong></summary>
-
-llamactl can be configured via configuration files or environment variables. Configuration is loaded in the following order of precedence:  
-
-```
-Defaults < Configuration file < Environment variables
-```
-
-### Configuration Files
-
-#### Configuration File Locations
-
-Configuration files are searched in the following locations (in order of precedence):
-
-**Linux/macOS:**
-- `./llamactl.yaml` or `./config.yaml` (current directory)
-- `$HOME/.config/llamactl/config.yaml`
-- `/etc/llamactl/config.yaml`
-
-**Windows:**
-- `./llamactl.yaml` or `./config.yaml` (current directory)
-- `%APPDATA%\llamactl\config.yaml`
-- `%USERPROFILE%\llamactl\config.yaml`
-- `%PROGRAMDATA%\llamactl\config.yaml`
-
-You can specify the path to config file with `LLAMACTL_CONFIG_PATH` environment variable.
-
-### Configuration Options
-
-#### Server Configuration
-
-```yaml
-server:
-  host: "0.0.0.0"         # Server host to bind to (default: "0.0.0.0")
-  port: 8080              # Server port to bind to (default: 8080)
-  allowed_origins: ["*"]  # CORS allowed origins (default: ["*"])
-  enable_swagger: false   # Enable Swagger UI (default: false)
-```
-
-**Environment Variables:**
-- `LLAMACTL_HOST` - Server host
-- `LLAMACTL_PORT` - Server port
-- `LLAMACTL_ALLOWED_ORIGINS` - Comma-separated CORS origins
-- `LLAMACTL_ENABLE_SWAGGER` - Enable Swagger UI (true/false)
-
-#### Instance Configuration
-
-```yaml
-instances:
-  port_range: [8000, 9000]                          # Port range for instances (default: [8000, 9000])
-  data_dir: "~/.local/share/llamactl"               # Directory for all llamactl data (default varies by OS)
-  configs_dir: "~/.local/share/llamactl/instances"  # Directory for instance configs (default: data_dir/instances)
-  logs_dir: "~/.local/share/llamactl/logs"          # Directory for instance logs (default: data_dir/logs)
-  auto_create_dirs: true                            # Automatically create data/config/logs directories (default: true)
-  max_instances: -1                                 # Maximum instances (-1 = unlimited)
-  llama_executable: "llama-server"                  # Path to llama-server executable
-  default_auto_restart: true                        # Default auto-restart setting
-  default_max_restarts: 3                           # Default maximum restart attempts
-  default_restart_delay: 5                          # Default restart delay in seconds
-  default_on_demand_start: true                     # Default on-demand start setting
-  on_demand_start_timeout: 120                      # Default on-demand start timeout in seconds
-  timeout_check_interval: 5                         # Default instance timeout check interval in minutes
-```
-
-**Environment Variables:**
-- `LLAMACTL_INSTANCE_PORT_RANGE` - Port range (format: "8000-9000" or "8000,9000")
-- `LLAMACTL_DATA_DIRECTORY` - Data directory path
-- `LLAMACTL_INSTANCES_DIR` - Instance configs directory path
-- `LLAMACTL_LOGS_DIR` - Log directory path
-- `LLAMACTL_AUTO_CREATE_DATA_DIR` - Auto-create data/config/logs directories (true/false)
-- `LLAMACTL_MAX_INSTANCES` - Maximum number of instances
-- `LLAMACTL_LLAMA_EXECUTABLE` - Path to llama-server executable
-- `LLAMACTL_DEFAULT_AUTO_RESTART` - Default auto-restart setting (true/false)
-- `LLAMACTL_DEFAULT_MAX_RESTARTS` - Default maximum restarts
-- `LLAMACTL_DEFAULT_RESTART_DELAY` - Default restart delay in seconds
-- `LLAMACTL_DEFAULT_ON_DEMAND_START` - Default on-demand start setting (true/false)
-- `LLAMACTL_ON_DEMAND_START_TIMEOUT` - Default on-demand start timeout in seconds
-- `LLAMACTL_TIMEOUT_CHECK_INTERVAL` - Default instance timeout check interval in minutes
-
-
-#### Authentication Configuration
-
-```yaml
-auth:
-  require_inference_auth: true           # Require API key for OpenAI endpoints (default: true)
-  inference_keys: []                     # List of valid inference API keys
-  require_management_auth: true          # Require API key for management endpoints (default: true)
-  management_keys: []                    # List of valid management API keys
-```
-
-**Environment Variables:**
-- `LLAMACTL_REQUIRE_INFERENCE_AUTH` - Require auth for OpenAI endpoints (true/false)
-- `LLAMACTL_INFERENCE_KEYS` - Comma-separated inference API keys
-- `LLAMACTL_REQUIRE_MANAGEMENT_AUTH` - Require auth for management endpoints (true/false)
-- `LLAMACTL_MANAGEMENT_KEYS` - Comma-separated management API keys
-
-</details>
+For detailed configuration options including environment variables, file locations, and advanced settings, see the [Configuration Guide](docs/getting-started/configuration.md).
 
 ## License
 

apidocs/docs.go

@@ -19,6 +19,243 @@ const docTemplate = `{
     "host": "{{.Host}}",
     "basePath": "{{.BasePath}}",
     "paths": {
+        "/backends/llama-cpp/devices": {
+            "get": {
+                "security": [
+                    {
+                        "ApiKeyAuth": []
+                    }
+                ],
+                "description": "Returns a list of available devices for the llama server",
+                "tags": [
+                    "backends"
+                ],
+                "summary": "List available devices for llama server",
+                "responses": {
+                    "200": {
+                        "description": "List of devices",
+                        "schema": {
+                            "type": "string"
+                        }
+                    },
+                    "500": {
+                        "description": "Internal Server Error",
+                        "schema": {
+                            "type": "string"
+                        }
+                    }
+                }
+            }
+        },
+        "/backends/llama-cpp/help": {
+            "get": {
+                "security": [
+                    {
+                        "ApiKeyAuth": []
+                    }
+                ],
+                "description": "Returns the help text for the llama server command",
+                "tags": [
+                    "backends"
+                ],
+                "summary": "Get help for llama server",
+                "responses": {
+                    "200": {
+                        "description": "Help text",
+                        "schema": {
+                            "type": "string"
+                        }
+                    },
+                    "500": {
+                        "description": "Internal Server Error",
+                        "schema": {
+                            "type": "string"
+                        }
+                    }
+                }
+            }
+        },
+        "/backends/llama-cpp/parse-command": {
+            "post": {
+                "security": [
+                    {
+                        "ApiKeyAuth": []
+                    }
+                ],
+                "description": "Parses a llama-server command string into instance options",
+                "consumes": [
+                    "application/json"
+                ],
+                "produces": [
+                    "application/json"
+                ],
+                "tags": [
+                    "backends"
+                ],
+                "summary": "Parse llama-server command",
+                "parameters": [
+                    {
+                        "description": "Command to parse",
+                        "name": "request",
+                        "in": "body",
+                        "required": true,
+                        "schema": {
+                            "$ref": "#/definitions/server.ParseCommandRequest"
+                        }
+                    }
+                ],
+                "responses": {
+                    "200": {
+                        "description": "Parsed options",
+                        "schema": {
+                            "$ref": "#/definitions/instance.CreateInstanceOptions"
+                        }
+                    },
+                    "400": {
+                        "description": "Invalid request or command",
+                        "schema": {
+                            "type": "object",
+                            "additionalProperties": {
+                                "type": "string"
+                            }
+                        }
+                    },
+                    "500": {
+                        "description": "Internal Server Error",
+                        "schema": {
+                            "type": "object",
+                            "additionalProperties": {
+                                "type": "string"
+                            }
+                        }
+                    }
+                }
+            }
+        },
+        "/backends/llama-cpp/version": {
+            "get": {
+                "security": [
+                    {
+                        "ApiKeyAuth": []
+                    }
+                ],
+                "description": "Returns the version of the llama server command",
+                "tags": [
+                    "backends"
+                ],
+                "summary": "Get version of llama server",
+                "responses": {
+                    "200": {
+                        "description": "Version information",
+                        "schema": {
+                            "type": "string"
+                        }
+                    },
+                    "500": {
+                        "description": "Internal Server Error",
+                        "schema": {
+                            "type": "string"
+                        }
+                    }
+                }
+            }
+        },
+        "/backends/mlx/parse-command": {
+            "post": {
+                "security": [
+                    {
+                        "ApiKeyAuth": []
+                    }
+                ],
+                "description": "Parses MLX-LM server command string into instance options",
+                "consumes": [
+                    "application/json"
+                ],
+                "produces": [
+                    "application/json"
+                ],
+                "tags": [
+                    "backends"
+                ],
+                "summary": "Parse mlx_lm.server command",
+                "parameters": [
+                    {
+                        "description": "Command to parse",
+                        "name": "request",
+                        "in": "body",
+                        "required": true,
+                        "schema": {
+                            "$ref": "#/definitions/server.ParseCommandRequest"
+                        }
+                    }
+                ],
+                "responses": {
+                    "200": {
+                        "description": "Parsed options",
+                        "schema": {
+                            "$ref": "#/definitions/instance.CreateInstanceOptions"
+                        }
+                    },
+                    "400": {
+                        "description": "Invalid request or command",
+                        "schema": {
+                            "type": "object",
+                            "additionalProperties": {
+                                "type": "string"
+                            }
+                        }
+                    }
+                }
+            }
+        },
+        "/backends/vllm/parse-command": {
+            "post": {
+                "security": [
+                    {
+                        "ApiKeyAuth": []
+                    }
+                ],
+                "description": "Parses a vLLM serve command string into instance options",
+                "consumes": [
+                    "application/json"
+                ],
+                "produces": [
+                    "application/json"
+                ],
+                "tags": [
+                    "backends"
+                ],
+                "summary": "Parse vllm serve command",
+                "parameters": [
+                    {
+                        "description": "Command to parse",
+                        "name": "request",
+                        "in": "body",
+                        "required": true,
+                        "schema": {
+                            "$ref": "#/definitions/server.ParseCommandRequest"
+                        }
+                    }
+                ],
+                "responses": {
+                    "200": {
+                        "description": "Parsed options",
+                        "schema": {
+                            "$ref": "#/definitions/instance.CreateInstanceOptions"
+                        }
+                    },
+                    "400": {
+                        "description": "Invalid request or command",
+                        "schema": {
+                            "type": "object",
+                            "additionalProperties": {
+                                "type": "string"
+                            }
+                        }
+                    }
+                }
+            }
+        },
         "/instances": {
             "get": {
                 "security": [
@@ -505,90 +742,6 @@ const docTemplate = `{
                 }
             }
         },
-        "/server/devices": {
-            "get": {
-                "security": [
-                    {
-                        "ApiKeyAuth": []
-                    }
-                ],
-                "description": "Returns a list of available devices for the llama server",
-                "tags": [
-                    "server"
-                ],
-                "summary": "List available devices for llama server",
-                "responses": {
-                    "200": {
-                        "description": "List of devices",
-                        "schema": {
-                            "type": "string"
-                        }
-                    },
-                    "500": {
-                        "description": "Internal Server Error",
-                        "schema": {
-                            "type": "string"
-                        }
-                    }
-                }
-            }
-        },
-        "/server/help": {
-            "get": {
-                "security": [
-                    {
-                        "ApiKeyAuth": []
-                    }
-                ],
-                "description": "Returns the help text for the llama server command",
-                "tags": [
-                    "server"
-                ],
-                "summary": "Get help for llama server",
-                "responses": {
-                    "200": {
-                        "description": "Help text",
-                        "schema": {
-                            "type": "string"
-                        }
-                    },
-                    "500": {
-                        "description": "Internal Server Error",
-                        "schema": {
-                            "type": "string"
-                        }
-                    }
-                }
-            }
-        },
-        "/server/version": {
-            "get": {
-                "security": [
-                    {
-                        "ApiKeyAuth": []
-                    }
-                ],
-                "description": "Returns the version of the llama server command",
-                "tags": [
-                    "server"
-                ],
-                "summary": "Get version of llama server",
-                "responses": {
-                    "200": {
-                        "description": "Version information",
-                        "schema": {
-                            "type": "string"
-                        }
-                    },
-                    "500": {
-                        "description": "Internal Server Error",
-                        "schema": {
-                            "type": "string"
-                        }
-                    }
-                }
-            }
-        },
         "/v1/": {
             "post": {
                 "security": [
@@ -609,7 +762,7 @@ const docTemplate = `{
                         "description": "OpenAI response"
                     },
                     "400": {
-                        "description": "Invalid request body or model name",
+                        "description": "Invalid request body or instance name",
                         "schema": {
                             "type": "string"
                         }
@@ -681,519 +834,63 @@ const docTemplate = `{
         }
     },
     "definitions": {
+        "backends.BackendType": {
+            "type": "string",
+            "enum": [
+                "llama_cpp",
+                "mlx_lm",
+                "vllm"
+            ],
+            "x-enum-varnames": [
+                "BackendTypeLlamaCpp",
+                "BackendTypeMlxLm",
+                "BackendTypeVllm"
+            ]
+        },
         "instance.CreateInstanceOptions": {
             "type": "object",
             "properties": {
-                "alias": {
-                    "type": "string"
-                },
-                "api_key": {
-                    "type": "string"
-                },
-                "api_key_file": {
-                    "type": "string"
-                },
                 "auto_restart": {
                     "description": "Auto restart",
                     "type": "boolean"
                 },
-                "batch_size": {
-                    "type": "integer"
+                "backend_options": {
+                    "type": "object",
+                    "additionalProperties": {}
                 },
-                "cache_reuse": {
-                    "type": "integer"
+                "backend_type": {
+                    "$ref": "#/definitions/backends.BackendType"
                 },
-                "cache_type_k": {
-                    "type": "string"
-                },
-                "cache_type_k_draft": {
-                    "type": "string"
-                },
-                "cache_type_v": {
-                    "type": "string"
-                },
-                "cache_type_v_draft": {
-                    "type": "string"
-                },
-                "chat_template": {
-                    "type": "string"
-                },
-                "chat_template_file": {
-                    "type": "string"
-                },
-                "chat_template_kwargs": {
-                    "type": "string"
-                },
-                "check_tensors": {
-                    "type": "boolean"
-                },
-                "cont_batching": {
-                    "type": "boolean"
-                },
-                "control_vector": {
-                    "type": "array",
-                    "items": {
-                        "type": "string"
-                    }
-                },
-                "control_vector_layer_range": {
-                    "type": "string"
-                },
-                "control_vector_scaled": {
-                    "type": "array",
-                    "items": {
-                        "type": "string"
-                    }
-                },
-                "cpu_mask": {
-                    "type": "string"
-                },
-                "cpu_mask_batch": {
-                    "type": "string"
-                },
-                "cpu_range": {
-                    "type": "string"
-                },
-                "cpu_range_batch": {
-                    "type": "string"
-                },
-                "cpu_strict": {
-                    "type": "integer"
-                },
-                "cpu_strict_batch": {
-                    "type": "integer"
-                },
-                "ctx_size": {
-                    "type": "integer"
-                },
-                "ctx_size_draft": {
-                    "type": "integer"
-                },
-                "defrag_thold": {
-                    "type": "number"
-                },
-                "device": {
-                    "type": "string"
-                },
-                "device_draft": {
-                    "type": "string"
-                },
-                "draft_max": {
-                    "type": "integer"
-                },
-                "draft_min": {
-                    "type": "integer"
-                },
-                "draft_p_min": {
-                    "type": "number"
-                },
-                "dry_allowed_length": {
-                    "type": "integer"
-                },
-                "dry_base": {
-                    "type": "number"
-                },
-                "dry_multiplier": {
-                    "type": "number"
-                },
-                "dry_penalty_last_n": {
-                    "type": "integer"
-                },
-                "dry_sequence_breaker": {
-                    "type": "array",
-                    "items": {
-                        "type": "string"
-                    }
-                },
-                "dump_kv_cache": {
-                    "type": "boolean"
-                },
-                "dynatemp_exp": {
-                    "type": "number"
-                },
-                "dynatemp_range": {
-                    "type": "number"
-                },
-                "embd_bge_small_en_default": {
-                    "description": "Default model params",
-                    "type": "boolean"
-                },
-                "embd_e5_small_en_default": {
-                    "type": "boolean"
-                },
-                "embd_gte_small_default": {
-                    "type": "boolean"
-                },
-                "embedding": {
-                    "type": "boolean"
-                },
-                "escape": {
-                    "type": "boolean"
-                },
-                "fim_qwen_14b_spec": {
-                    "type": "boolean"
-                },
-                "fim_qwen_1_5b_default": {
-                    "type": "boolean"
-                },
-                "fim_qwen_3b_default": {
-                    "type": "boolean"
-                },
-                "fim_qwen_7b_default": {
-                    "type": "boolean"
-                },
-                "fim_qwen_7b_spec": {
-                    "type": "boolean"
-                },
-                "flash_attn": {
-                    "type": "boolean"
-                },
-                "frequency_penalty": {
-                    "type": "number"
-                },
-                "gpu_layers": {
-                    "type": "integer"
-                },
-                "gpu_layers_draft": {
-                    "type": "integer"
-                },
-                "grammar": {
-                    "type": "string"
-                },
-                "grammar_file": {
-                    "type": "string"
-                },
-                "hf_file": {
-                    "type": "string"
-                },
-                "hf_file_v": {
-                    "type": "string"
-                },
-                "hf_repo": {
-                    "type": "string"
-                },
-                "hf_repo_draft": {
-                    "type": "string"
-                },
-                "hf_repo_v": {
-                    "type": "string"
-                },
-                "hf_token": {
-                    "type": "string"
-                },
-                "host": {
-                    "type": "string"
-                },
-                "ignore_eos": {
-                    "type": "boolean"
-                },
-                "jinja": {
-                    "type": "boolean"
-                },
-                "json_schema": {
-                    "type": "string"
-                },
-                "json_schema_file": {
-                    "type": "string"
-                },
-                "keep": {
-                    "type": "integer"
-                },
-                "log_colors": {
-                    "type": "boolean"
-                },
-                "log_disable": {
-                    "type": "boolean"
-                },
-                "log_file": {
-                    "type": "string"
-                },
-                "log_prefix": {
-                    "type": "boolean"
-                },
-                "log_timestamps": {
-                    "type": "boolean"
-                },
-                "logit_bias": {
-                    "type": "array",
-                    "items": {
-                        "type": "string"
-                    }
-                },
-                "lora": {
-                    "type": "array",
-                    "items": {
-                        "type": "string"
-                    }
-                },
-                "lora_init_without_apply": {
-                    "type": "boolean"
-                },
-                "lora_scaled": {
-                    "type": "array",
-                    "items": {
-                        "type": "string"
-                    }
-                },
-                "main_gpu": {
+                "idle_timeout": {
+                    "description": "Idle timeout",
                     "type": "integer"
                 },
                 "max_restarts": {
                     "type": "integer"
                 },
-                "metrics": {
+                "on_demand_start": {
+                    "description": "On demand start",
                     "type": "boolean"
                 },
-                "min_p": {
-                    "type": "number"
-                },
-                "mirostat": {
-                    "type": "integer"
-                },
-                "mirostat_ent": {
-                    "type": "number"
-                },
-                "mirostat_lr": {
-                    "type": "number"
-                },
-                "mlock": {
-                    "type": "boolean"
-                },
-                "mmproj": {
-                    "type": "string"
-                },
-                "mmproj_url": {
-                    "type": "string"
-                },
-                "model": {
-                    "type": "string"
-                },
-                "model_draft": {
-                    "type": "string"
-                },
-                "model_url": {
-                    "type": "string"
-                },
-                "model_vocoder": {
-                    "description": "Audio/TTS params",
-                    "type": "string"
-                },
-                "no_cont_batching": {
-                    "type": "boolean"
-                },
-                "no_context_shift": {
-                    "description": "Example-specific params",
-                    "type": "boolean"
-                },
-                "no_escape": {
-                    "type": "boolean"
-                },
-                "no_kv_offload": {
-                    "type": "boolean"
-                },
-                "no_mmap": {
-                    "type": "boolean"
-                },
-                "no_mmproj": {
-                    "type": "boolean"
-                },
-                "no_mmproj_offload": {
-                    "type": "boolean"
-                },
-                "no_perf": {
-                    "type": "boolean"
-                },
-                "no_prefill_assistant": {
-                    "type": "boolean"
-                },
-                "no_slots": {
-                    "type": "boolean"
-                },
-                "no_warmup": {
-                    "type": "boolean"
-                },
-                "no_webui": {
-                    "type": "boolean"
-                },
-                "numa": {
-                    "type": "string"
-                },
-                "override_kv": {
-                    "type": "array",
-                    "items": {
-                        "type": "string"
-                    }
-                },
-                "override_tensor": {
-                    "type": "array",
-                    "items": {
-                        "type": "string"
-                    }
-                },
-                "parallel": {
-                    "type": "integer"
-                },
-                "path": {
-                    "type": "string"
-                },
-                "poll": {
-                    "type": "integer"
-                },
-                "poll_batch": {
-                    "type": "integer"
-                },
-                "pooling": {
-                    "type": "string"
-                },
-                "port": {
-                    "type": "integer"
-                },
-                "predict": {
-                    "type": "integer"
-                },
-                "presence_penalty": {
-                    "type": "number"
-                },
-                "prio": {
-                    "type": "integer"
-                },
-                "prio_batch": {
-                    "type": "integer"
-                },
-                "props": {
-                    "type": "boolean"
-                },
-                "reasoning_budget": {
-                    "type": "integer"
-                },
-                "reasoning_format": {
-                    "type": "string"
-                },
-                "repeat_last_n": {
-                    "type": "integer"
-                },
-                "repeat_penalty": {
-                    "type": "number"
-                },
-                "reranking": {
-                    "type": "boolean"
-                },
-                "restart_delay_seconds": {
-                    "description": "RestartDelay duration in seconds",
-                    "type": "integer"
-                },
-                "rope_freq_base": {
-                    "type": "number"
-                },
-                "rope_freq_scale": {
-                    "type": "number"
-                },
-                "rope_scale": {
-                    "type": "number"
-                },
-                "rope_scaling": {
-                    "type": "string"
-                },
-                "samplers": {
-                    "description": "Sampling params",
-                    "type": "string"
-                },
-                "sampling_seq": {
-                    "type": "string"
-                },
-                "seed": {
-                    "type": "integer"
-                },
-                "slot_prompt_similarity": {
-                    "type": "number"
-                },
-                "slot_save_path": {
-                    "type": "string"
-                },
-                "slots": {
-                    "type": "boolean"
-                },
-                "special": {
-                    "type": "boolean"
-                },
-                "split_mode": {
-                    "type": "string"
-                },
-                "spm_infill": {
-                    "type": "boolean"
-                },
-                "ssl_cert_file": {
-                    "type": "string"
-                },
-                "ssl_key_file": {
-                    "type": "string"
-                },
-                "temp": {
-                    "type": "number"
-                },
-                "tensor_split": {
-                    "type": "string"
-                },
-                "threads": {
-                    "type": "integer"
-                },
-                "threads_batch": {
-                    "type": "integer"
-                },
-                "threads_http": {
-                    "type": "integer"
-                },
-                "timeout": {
-                    "type": "integer"
-                },
-                "top_k": {
-                    "type": "integer"
-                },
-                "top_p": {
-                    "type": "number"
-                },
-                "tts_use_guide_tokens": {
-                    "type": "boolean"
-                },
-                "typical": {
-                    "type": "number"
-                },
-                "ubatch_size": {
-                    "type": "integer"
-                },
-                "verbose": {
-                    "type": "boolean"
-                },
-                "verbose_prompt": {
-                    "description": "Common params",
-                    "type": "boolean"
-                },
-                "verbosity": {
-                    "type": "integer"
-                },
-                "xtc_probability": {
-                    "type": "number"
-                },
-                "xtc_threshold": {
-                    "type": "number"
-                },
-                "yarn_attn_factor": {
-                    "type": "number"
-                },
-                "yarn_beta_fast": {
-                    "type": "number"
-                },
-                "yarn_beta_slow": {
-                    "type": "number"
-                },
-                "yarn_ext_factor": {
-                    "type": "number"
-                },
-                "yarn_orig_ctx": {
+                "restart_delay": {
+                    "description": "seconds",
                     "type": "integer"
                 }
             }
         },
+        "instance.InstanceStatus": {
+            "type": "integer",
+            "enum": [
+                0,
+                1,
+                2
+            ],
+            "x-enum-varnames": [
+                "Stopped",
+                "Running",
+                "Failed"
+            ]
+        },
         "instance.Process": {
             "type": "object",
             "properties": {
@@ -1204,9 +901,13 @@ const docTemplate = `{
                 "name": {
                     "type": "string"
                 },
-                "running": {
+                "status": {
                     "description": "Status",
-                    "type": "boolean"
+                    "allOf": [
+                        {
+                            "$ref": "#/definitions/instance.InstanceStatus"
+                        }
+                    ]
                 }
             }
         },
@@ -1240,6 +941,14 @@ const docTemplate = `{
                     "type": "string"
                 }
             }
+        },
+        "server.ParseCommandRequest": {
+            "type": "object",
+            "properties": {
+                "command": {
+                    "type": "string"
+                }
+            }
         }
     }
 }`

apidocs/swagger.json

@@ -12,6 +12,243 @@
     },
     "basePath": "/api/v1",
     "paths": {
+        "/backends/llama-cpp/devices": {
+            "get": {
+                "security": [
+                    {
+                        "ApiKeyAuth": []
+                    }
+                ],
+                "description": "Returns a list of available devices for the llama server",
+                "tags": [
+                    "backends"
+                ],
+                "summary": "List available devices for llama server",
+                "responses": {
+                    "200": {
+                        "description": "List of devices",
+                        "schema": {
+                            "type": "string"
+                        }
+                    },
+                    "500": {
+                        "description": "Internal Server Error",
+                        "schema": {
+                            "type": "string"
+                        }
+                    }
+                }
+            }
+        },
+        "/backends/llama-cpp/help": {
+            "get": {
+                "security": [
+                    {
+                        "ApiKeyAuth": []
+                    }
+                ],
+                "description": "Returns the help text for the llama server command",
+                "tags": [
+                    "backends"
+                ],
+                "summary": "Get help for llama server",
+                "responses": {
+                    "200": {
+                        "description": "Help text",
+                        "schema": {
+                            "type": "string"
+                        }
+                    },
+                    "500": {
+                        "description": "Internal Server Error",
+                        "schema": {
+                            "type": "string"
+                        }
+                    }
+                }
+            }
+        },
+        "/backends/llama-cpp/parse-command": {
+            "post": {
+                "security": [
+                    {
+                        "ApiKeyAuth": []
+                    }
+                ],
+                "description": "Parses a llama-server command string into instance options",
+                "consumes": [
+                    "application/json"
+                ],
+                "produces": [
+                    "application/json"
+                ],
+                "tags": [
+                    "backends"
+                ],
+                "summary": "Parse llama-server command",
+                "parameters": [
+                    {
+                        "description": "Command to parse",
+                        "name": "request",
+                        "in": "body",
+                        "required": true,
+                        "schema": {
+                            "$ref": "#/definitions/server.ParseCommandRequest"
+                        }
+                    }
+                ],
+                "responses": {
+                    "200": {
+                        "description": "Parsed options",
+                        "schema": {
+                            "$ref": "#/definitions/instance.CreateInstanceOptions"
+                        }
+                    },
+                    "400": {
+                        "description": "Invalid request or command",
+                        "schema": {
+                            "type": "object",
+                            "additionalProperties": {
+                                "type": "string"
+                            }
+                        }
+                    },
+                    "500": {
+                        "description": "Internal Server Error",
+                        "schema": {
+                            "type": "object",
+                            "additionalProperties": {
+                                "type": "string"
+                            }
+                        }
+                    }
+                }
+            }
+        },
+        "/backends/llama-cpp/version": {
+            "get": {
+                "security": [
+                    {
+                        "ApiKeyAuth": []
+                    }
+                ],
+                "description": "Returns the version of the llama server command",
+                "tags": [
+                    "backends"
+                ],
+                "summary": "Get version of llama server",
+                "responses": {
+                    "200": {
+                        "description": "Version information",
+                        "schema": {
+                            "type": "string"
+                        }
+                    },
+                    "500": {
+                        "description": "Internal Server Error",
+                        "schema": {
+                            "type": "string"
+                        }
+                    }
+                }
+            }
+        },
+        "/backends/mlx/parse-command": {
+            "post": {
+                "security": [
+                    {
+                        "ApiKeyAuth": []
+                    }
+                ],
+                "description": "Parses MLX-LM server command string into instance options",
+                "consumes": [
+                    "application/json"
+                ],
+                "produces": [
+                    "application/json"
+                ],
+                "tags": [
+                    "backends"
+                ],
+                "summary": "Parse mlx_lm.server command",
+                "parameters": [
+                    {
+                        "description": "Command to parse",
+                        "name": "request",
+                        "in": "body",
+                        "required": true,
+                        "schema": {
+                            "$ref": "#/definitions/server.ParseCommandRequest"
+                        }
+                    }
+                ],
+                "responses": {
+                    "200": {
+                        "description": "Parsed options",
+                        "schema": {
+                            "$ref": "#/definitions/instance.CreateInstanceOptions"
+                        }
+                    },
+                    "400": {
+                        "description": "Invalid request or command",
+                        "schema": {
+                            "type": "object",
+                            "additionalProperties": {
+                                "type": "string"
+                            }
+                        }
+                    }
+                }
+            }
+        },
+        "/backends/vllm/parse-command": {
+            "post": {
+                "security": [
+                    {
+                        "ApiKeyAuth": []
+                    }
+                ],
+                "description": "Parses a vLLM serve command string into instance options",
+                "consumes": [
+                    "application/json"
+                ],
+                "produces": [
+                    "application/json"
+                ],
+                "tags": [
+                    "backends"
+                ],
+                "summary": "Parse vllm serve command",
+                "parameters": [
+                    {
+                        "description": "Command to parse",
+                        "name": "request",
+                        "in": "body",
+                        "required": true,
+                        "schema": {
+                            "$ref": "#/definitions/server.ParseCommandRequest"
+                        }
+                    }
+                ],
+                "responses": {
+                    "200": {
+                        "description": "Parsed options",
+                        "schema": {
+                            "$ref": "#/definitions/instance.CreateInstanceOptions"
+                        }
+                    },
+                    "400": {
+                        "description": "Invalid request or command",
+                        "schema": {
+                            "type": "object",
+                            "additionalProperties": {
+                                "type": "string"
+                            }
+                        }
+                    }
+                }
+            }
+        },
         "/instances": {
             "get": {
                 "security": [
@@ -498,90 +735,6 @@
                 }
             }
         },
-        "/server/devices": {
-            "get": {
-                "security": [
-                    {
-                        "ApiKeyAuth": []
-                    }
-                ],
-                "description": "Returns a list of available devices for the llama server",
-                "tags": [
-                    "server"
-                ],
-                "summary": "List available devices for llama server",
-                "responses": {
-                    "200": {
-                        "description": "List of devices",
-                        "schema": {
-                            "type": "string"
-                        }
-                    },
-                    "500": {
-                        "description": "Internal Server Error",
-                        "schema": {
-                            "type": "string"
-                        }
-                    }
-                }
-            }
-        },
-        "/server/help": {
-            "get": {
-                "security": [
-                    {
-                        "ApiKeyAuth": []
-                    }
-                ],
-                "description": "Returns the help text for the llama server command",
-                "tags": [
-                    "server"
-                ],
-                "summary": "Get help for llama server",
-                "responses": {
-                    "200": {
-                        "description": "Help text",
-                        "schema": {
-                            "type": "string"
-                        }
-                    },
-                    "500": {
-                        "description": "Internal Server Error",
-                        "schema": {
-                            "type": "string"
-                        }
-                    }
-                }
-            }
-        },
-        "/server/version": {
-            "get": {
-                "security": [
-                    {
-                        "ApiKeyAuth": []
-                    }
-                ],
-                "description": "Returns the version of the llama server command",
-                "tags": [
-                    "server"
-                ],
-                "summary": "Get version of llama server",
-                "responses": {
-                    "200": {
-                        "description": "Version information",
-                        "schema": {
-                            "type": "string"
-                        }
-                    },
-                    "500": {
-                        "description": "Internal Server Error",
-                        "schema": {
-                            "type": "string"
-                        }
-                    }
-                }
-            }
-        },
         "/v1/": {
             "post": {
                 "security": [
@@ -602,7 +755,7 @@
                         "description": "OpenAI response"
                     },
                     "400": {
-                        "description": "Invalid request body or model name",
+                        "description": "Invalid request body or instance name",
                         "schema": {
                             "type": "string"
                         }
@@ -674,519 +827,63 @@
         }
     },
     "definitions": {
+        "backends.BackendType": {
+            "type": "string",
+            "enum": [
+                "llama_cpp",
+                "mlx_lm",
+                "vllm"
+            ],
+            "x-enum-varnames": [
+                "BackendTypeLlamaCpp",
+                "BackendTypeMlxLm",
+                "BackendTypeVllm"
+            ]
+        },
         "instance.CreateInstanceOptions": {
             "type": "object",
             "properties": {
-                "alias": {
-                    "type": "string"
-                },
-                "api_key": {
-                    "type": "string"
-                },
-                "api_key_file": {
-                    "type": "string"
-                },
                 "auto_restart": {
                     "description": "Auto restart",
                     "type": "boolean"
                 },
-                "batch_size": {
-                    "type": "integer"
+                "backend_options": {
+                    "type": "object",
+                    "additionalProperties": {}
                 },
-                "cache_reuse": {
-                    "type": "integer"
+                "backend_type": {
+                    "$ref": "#/definitions/backends.BackendType"
                 },
-                "cache_type_k": {
-                    "type": "string"
-                },
-                "cache_type_k_draft": {
-                    "type": "string"
-                },
-                "cache_type_v": {
-                    "type": "string"
-                },
-                "cache_type_v_draft": {
-                    "type": "string"
-                },
-                "chat_template": {
-                    "type": "string"
-                },
-                "chat_template_file": {
-                    "type": "string"
-                },
-                "chat_template_kwargs": {
-                    "type": "string"
-                },
-                "check_tensors": {
-                    "type": "boolean"
-                },
-                "cont_batching": {
-                    "type": "boolean"
-                },
-                "control_vector": {
-                    "type": "array",
-                    "items": {
-                        "type": "string"
-                    }
-                },
-                "control_vector_layer_range": {
-                    "type": "string"
-                },
-                "control_vector_scaled": {
-                    "type": "array",
-                    "items": {
-                        "type": "string"
-                    }
-                },
-                "cpu_mask": {
-                    "type": "string"
-                },
-                "cpu_mask_batch": {
-                    "type": "string"
-                },
-                "cpu_range": {
-                    "type": "string"
-                },
-                "cpu_range_batch": {
-                    "type": "string"
-                },
-                "cpu_strict": {
-                    "type": "integer"
-                },
-                "cpu_strict_batch": {
-                    "type": "integer"
-                },
-                "ctx_size": {
-                    "type": "integer"
-                },
-                "ctx_size_draft": {
-                    "type": "integer"
-                },
-                "defrag_thold": {
-                    "type": "number"
-                },
-                "device": {
-                    "type": "string"
-                },
-                "device_draft": {
-                    "type": "string"
-                },
-                "draft_max": {
-                    "type": "integer"
-                },
-                "draft_min": {
-                    "type": "integer"
-                },
-                "draft_p_min": {
-                    "type": "number"
-                },
-                "dry_allowed_length": {
-                    "type": "integer"
-                },
-                "dry_base": {
-                    "type": "number"
-                },
-                "dry_multiplier": {
-                    "type": "number"
-                },
-                "dry_penalty_last_n": {
-                    "type": "integer"
-                },
-                "dry_sequence_breaker": {
-                    "type": "array",
-                    "items": {
-                        "type": "string"
-                    }
-                },
-                "dump_kv_cache": {
-                    "type": "boolean"
-                },
-                "dynatemp_exp": {
-                    "type": "number"
-                },
-                "dynatemp_range": {
-                    "type": "number"
-                },
-                "embd_bge_small_en_default": {
-                    "description": "Default model params",
-                    "type": "boolean"
-                },
-                "embd_e5_small_en_default": {
-                    "type": "boolean"
-                },
-                "embd_gte_small_default": {
-                    "type": "boolean"
-                },
-                "embedding": {
-                    "type": "boolean"
-                },
-                "escape": {
-                    "type": "boolean"
-                },
-                "fim_qwen_14b_spec": {
-                    "type": "boolean"
-                },
-                "fim_qwen_1_5b_default": {
-                    "type": "boolean"
-                },
-                "fim_qwen_3b_default": {
-                    "type": "boolean"
-                },
-                "fim_qwen_7b_default": {
-                    "type": "boolean"
-                },
-                "fim_qwen_7b_spec": {
-                    "type": "boolean"
-                },
-                "flash_attn": {
-                    "type": "boolean"
-                },
-                "frequency_penalty": {
-                    "type": "number"
-                },
-                "gpu_layers": {
-                    "type": "integer"
-                },
-                "gpu_layers_draft": {
-                    "type": "integer"
-                },
-                "grammar": {
-                    "type": "string"
-                },
-                "grammar_file": {
-                    "type": "string"
-                },
-                "hf_file": {
-                    "type": "string"
-                },
-                "hf_file_v": {
-                    "type": "string"
-                },
-                "hf_repo": {
-                    "type": "string"
-                },
-                "hf_repo_draft": {
-                    "type": "string"
-                },
-                "hf_repo_v": {
-                    "type": "string"
-                },
-                "hf_token": {
-                    "type": "string"
-                },
-                "host": {
-                    "type": "string"
-                },
-                "ignore_eos": {
-                    "type": "boolean"
-                },
-                "jinja": {
-                    "type": "boolean"
-                },
-                "json_schema": {
-                    "type": "string"
-                },
-                "json_schema_file": {
-                    "type": "string"
-                },
-                "keep": {
-                    "type": "integer"
-                },
-                "log_colors": {
-                    "type": "boolean"
-                },
-                "log_disable": {
-                    "type": "boolean"
-                },
-                "log_file": {
-                    "type": "string"
-                },
-                "log_prefix": {
-                    "type": "boolean"
-                },
-                "log_timestamps": {
-                    "type": "boolean"
-                },
-                "logit_bias": {
-                    "type": "array",
-                    "items": {
-                        "type": "string"
-                    }
-                },
-                "lora": {
-                    "type": "array",
-                    "items": {
-                        "type": "string"
-                    }
-                },
-                "lora_init_without_apply": {
-                    "type": "boolean"
-                },
-                "lora_scaled": {
-                    "type": "array",
-                    "items": {
-                        "type": "string"
-                    }
-                },
-                "main_gpu": {
+                "idle_timeout": {
+                    "description": "Idle timeout",
                     "type": "integer"
                 },
                 "max_restarts": {
                     "type": "integer"
                 },
-                "metrics": {
+                "on_demand_start": {
+                    "description": "On demand start",
                     "type": "boolean"
                 },
-                "min_p": {
-                    "type": "number"
-                },
-                "mirostat": {
-                    "type": "integer"
-                },
-                "mirostat_ent": {
-                    "type": "number"
-                },
-                "mirostat_lr": {
-                    "type": "number"
-                },
-                "mlock": {
-                    "type": "boolean"
-                },
-                "mmproj": {
-                    "type": "string"
-                },
-                "mmproj_url": {
-                    "type": "string"
-                },
-                "model": {
-                    "type": "string"
-                },
-                "model_draft": {
-                    "type": "string"
-                },
-                "model_url": {
-                    "type": "string"
-                },
-                "model_vocoder": {
-                    "description": "Audio/TTS params",
-                    "type": "string"
-                },
-                "no_cont_batching": {
-                    "type": "boolean"
-                },
-                "no_context_shift": {
-                    "description": "Example-specific params",
-                    "type": "boolean"
-                },
-                "no_escape": {
-                    "type": "boolean"
-                },
-                "no_kv_offload": {
-                    "type": "boolean"
-                },
-                "no_mmap": {
-                    "type": "boolean"
-                },
-                "no_mmproj": {
-                    "type": "boolean"
-                },
-                "no_mmproj_offload": {
-                    "type": "boolean"
-                },
-                "no_perf": {
-                    "type": "boolean"
-                },
-                "no_prefill_assistant": {
-                    "type": "boolean"
-                },
-                "no_slots": {
-                    "type": "boolean"
-                },
-                "no_warmup": {
-                    "type": "boolean"
-                },
-                "no_webui": {
-                    "type": "boolean"
-                },
-                "numa": {
-                    "type": "string"
-                },
-                "override_kv": {
-                    "type": "array",
-                    "items": {
-                        "type": "string"
-                    }
-                },
-                "override_tensor": {
-                    "type": "array",
-                    "items": {
-                        "type": "string"
-                    }
-                },
-                "parallel": {
-                    "type": "integer"
-                },
-                "path": {
-                    "type": "string"
-                },
-                "poll": {
-                    "type": "integer"
-                },
-                "poll_batch": {
-                    "type": "integer"
-                },
-                "pooling": {
-                    "type": "string"
-                },
-                "port": {
-                    "type": "integer"
-                },
-                "predict": {
-                    "type": "integer"
-                },
-                "presence_penalty": {
-                    "type": "number"
-                },
-                "prio": {
-                    "type": "integer"
-                },
-                "prio_batch": {
-                    "type": "integer"
-                },
-                "props": {
-                    "type": "boolean"
-                },
-                "reasoning_budget": {
-                    "type": "integer"
-                },
-                "reasoning_format": {
-                    "type": "string"
-                },
-                "repeat_last_n": {
-                    "type": "integer"
-                },
-                "repeat_penalty": {
-                    "type": "number"
-                },
-                "reranking": {
-                    "type": "boolean"
-                },
-                "restart_delay_seconds": {
-                    "description": "RestartDelay duration in seconds",
-                    "type": "integer"
-                },
-                "rope_freq_base": {
-                    "type": "number"
-                },
-                "rope_freq_scale": {
-                    "type": "number"
-                },
-                "rope_scale": {
-                    "type": "number"
-                },
-                "rope_scaling": {
-                    "type": "string"
-                },
-                "samplers": {
-                    "description": "Sampling params",
-                    "type": "string"
-                },
-                "sampling_seq": {
-                    "type": "string"
-                },
-                "seed": {
-                    "type": "integer"
-                },
-                "slot_prompt_similarity": {
-                    "type": "number"
-                },
-                "slot_save_path": {
-                    "type": "string"
-                },
-                "slots": {
-                    "type": "boolean"
-                },
-                "special": {
-                    "type": "boolean"
-                },
-                "split_mode": {
-                    "type": "string"
-                },
-                "spm_infill": {
-                    "type": "boolean"
-                },
-                "ssl_cert_file": {
-                    "type": "string"
-                },
-                "ssl_key_file": {
-                    "type": "string"
-                },
-                "temp": {
-                    "type": "number"
-                },
-                "tensor_split": {
-                    "type": "string"
-                },
-                "threads": {
-                    "type": "integer"
-                },
-                "threads_batch": {
-                    "type": "integer"
-                },
-                "threads_http": {
-                    "type": "integer"
-                },
-                "timeout": {
-                    "type": "integer"
-                },
-                "top_k": {
-                    "type": "integer"
-                },
-                "top_p": {
-                    "type": "number"
-                },
-                "tts_use_guide_tokens": {
-                    "type": "boolean"
-                },
-                "typical": {
-                    "type": "number"
-                },
-                "ubatch_size": {
-                    "type": "integer"
-                },
-                "verbose": {
-                    "type": "boolean"
-                },
-                "verbose_prompt": {
-                    "description": "Common params",
-                    "type": "boolean"
-                },
-                "verbosity": {
-                    "type": "integer"
-                },
-                "xtc_probability": {
-                    "type": "number"
-                },
-                "xtc_threshold": {
-                    "type": "number"
-                },
-                "yarn_attn_factor": {
-                    "type": "number"
-                },
-                "yarn_beta_fast": {
-                    "type": "number"
-                },
-                "yarn_beta_slow": {
-                    "type": "number"
-                },
-                "yarn_ext_factor": {
-                    "type": "number"
-                },
-                "yarn_orig_ctx": {
+                "restart_delay": {
+                    "description": "seconds",
                     "type": "integer"
                 }
             }
         },
+        "instance.InstanceStatus": {
+            "type": "integer",
+            "enum": [
+                0,
+                1,
+                2
+            ],
+            "x-enum-varnames": [
+                "Stopped",
+                "Running",
+                "Failed"
+            ]
+        },
         "instance.Process": {
             "type": "object",
             "properties": {
@@ -1197,9 +894,13 @@
                 "name": {
                     "type": "string"
                 },
-                "running": {
+                "status": {
                     "description": "Status",
-                    "type": "boolean"
+                    "allOf": [
+                        {
+                            "$ref": "#/definitions/instance.InstanceStatus"
+                        }
+                    ]
                 }
             }
         },
@@ -1233,6 +934,14 @@
                     "type": "string"
                 }
             }
+        },
+        "server.ParseCommandRequest": {
+            "type": "object",
+            "properties": {
+                "command": {
+                    "type": "string"
+                }
+            }
         }
     }
 }

apidocs/swagger.yaml

@@ -1,349 +1,47 @@
 basePath: /api/v1
 definitions:
+  backends.BackendType:
+    enum:
+    - llama_cpp
+    - mlx_lm
+    - vllm
+    type: string
+    x-enum-varnames:
+    - BackendTypeLlamaCpp
+    - BackendTypeMlxLm
+    - BackendTypeVllm
   instance.CreateInstanceOptions:
     properties:
-      alias:
-        type: string
-      api_key:
-        type: string
-      api_key_file:
-        type: string
       auto_restart:
         description: Auto restart
         type: boolean
-      batch_size:
-        type: integer
-      cache_reuse:
-        type: integer
-      cache_type_k:
-        type: string
-      cache_type_k_draft:
-        type: string
-      cache_type_v:
-        type: string
-      cache_type_v_draft:
-        type: string
-      chat_template:
-        type: string
-      chat_template_file:
-        type: string
-      chat_template_kwargs:
-        type: string
-      check_tensors:
-        type: boolean
-      cont_batching:
-        type: boolean
-      control_vector:
-        items:
-          type: string
-        type: array
-      control_vector_layer_range:
-        type: string
-      control_vector_scaled:
-        items:
-          type: string
-        type: array
-      cpu_mask:
-        type: string
-      cpu_mask_batch:
-        type: string
-      cpu_range:
-        type: string
-      cpu_range_batch:
-        type: string
-      cpu_strict:
-        type: integer
-      cpu_strict_batch:
-        type: integer
-      ctx_size:
-        type: integer
-      ctx_size_draft:
-        type: integer
-      defrag_thold:
-        type: number
-      device:
-        type: string
-      device_draft:
-        type: string
-      draft_max:
-        type: integer
-      draft_min:
-        type: integer
-      draft_p_min:
-        type: number
-      dry_allowed_length:
-        type: integer
-      dry_base:
-        type: number
-      dry_multiplier:
-        type: number
-      dry_penalty_last_n:
-        type: integer
-      dry_sequence_breaker:
-        items:
-          type: string
-        type: array
-      dump_kv_cache:
-        type: boolean
-      dynatemp_exp:
-        type: number
-      dynatemp_range:
-        type: number
-      embd_bge_small_en_default:
-        description: Default model params
-        type: boolean
-      embd_e5_small_en_default:
-        type: boolean
-      embd_gte_small_default:
-        type: boolean
-      embedding:
-        type: boolean
-      escape:
-        type: boolean
-      fim_qwen_1_5b_default:
-        type: boolean
-      fim_qwen_3b_default:
-        type: boolean
-      fim_qwen_7b_default:
-        type: boolean
-      fim_qwen_7b_spec:
-        type: boolean
-      fim_qwen_14b_spec:
-        type: boolean
-      flash_attn:
-        type: boolean
-      frequency_penalty:
-        type: number
-      gpu_layers:
-        type: integer
-      gpu_layers_draft:
-        type: integer
-      grammar:
-        type: string
-      grammar_file:
-        type: string
-      hf_file:
-        type: string
-      hf_file_v:
-        type: string
-      hf_repo:
-        type: string
-      hf_repo_draft:
-        type: string
-      hf_repo_v:
-        type: string
-      hf_token:
-        type: string
-      host:
-        type: string
-      ignore_eos:
-        type: boolean
-      jinja:
-        type: boolean
-      json_schema:
-        type: string
-      json_schema_file:
-        type: string
-      keep:
-        type: integer
-      log_colors:
-        type: boolean
-      log_disable:
-        type: boolean
-      log_file:
-        type: string
-      log_prefix:
-        type: boolean
-      log_timestamps:
-        type: boolean
-      logit_bias:
-        items:
-          type: string
-        type: array
-      lora:
-        items:
-          type: string
-        type: array
-      lora_init_without_apply:
-        type: boolean
-      lora_scaled:
-        items:
-          type: string
-        type: array
-      main_gpu:
+      backend_options:
+        additionalProperties: {}
+        type: object
+      backend_type:
+        $ref: '#/definitions/backends.BackendType'
+      idle_timeout:
+        description: Idle timeout
         type: integer
       max_restarts:
         type: integer
-      metrics:
+      on_demand_start:
+        description: On demand start
         type: boolean
-      min_p:
-        type: number
-      mirostat:
-        type: integer
-      mirostat_ent:
-        type: number
-      mirostat_lr:
-        type: number
-      mlock:
-        type: boolean
-      mmproj:
-        type: string
-      mmproj_url:
-        type: string
-      model:
-        type: string
-      model_draft:
-        type: string
-      model_url:
-        type: string
-      model_vocoder:
-        description: Audio/TTS params
-        type: string
-      no_cont_batching:
-        type: boolean
-      no_context_shift:
-        description: Example-specific params
-        type: boolean
-      no_escape:
-        type: boolean
-      no_kv_offload:
-        type: boolean
-      no_mmap:
-        type: boolean
-      no_mmproj:
-        type: boolean
-      no_mmproj_offload:
-        type: boolean
-      no_perf:
-        type: boolean
-      no_prefill_assistant:
-        type: boolean
-      no_slots:
-        type: boolean
-      no_warmup:
-        type: boolean
-      no_webui:
-        type: boolean
-      numa:
-        type: string
-      override_kv:
-        items:
-          type: string
-        type: array
-      override_tensor:
-        items:
-          type: string
-        type: array
-      parallel:
-        type: integer
-      path:
-        type: string
-      poll:
-        type: integer
-      poll_batch:
-        type: integer
-      pooling:
-        type: string
-      port:
-        type: integer
-      predict:
-        type: integer
-      presence_penalty:
-        type: number
-      prio:
-        type: integer
-      prio_batch:
-        type: integer
-      props:
-        type: boolean
-      reasoning_budget:
-        type: integer
-      reasoning_format:
-        type: string
-      repeat_last_n:
-        type: integer
-      repeat_penalty:
-        type: number
-      reranking:
-        type: boolean
-      restart_delay_seconds:
-        description: RestartDelay duration in seconds
-        type: integer
-      rope_freq_base:
-        type: number
-      rope_freq_scale:
-        type: number
-      rope_scale:
-        type: number
-      rope_scaling:
-        type: string
-      samplers:
-        description: Sampling params
-        type: string
-      sampling_seq:
-        type: string
-      seed:
-        type: integer
-      slot_prompt_similarity:
-        type: number
-      slot_save_path:
-        type: string
-      slots:
-        type: boolean
-      special:
-        type: boolean
-      split_mode:
-        type: string
-      spm_infill:
-        type: boolean
-      ssl_cert_file:
-        type: string
-      ssl_key_file:
-        type: string
-      temp:
-        type: number
-      tensor_split:
-        type: string
-      threads:
-        type: integer
-      threads_batch:
-        type: integer
-      threads_http:
-        type: integer
-      timeout:
-        type: integer
-      top_k:
-        type: integer
-      top_p:
-        type: number
-      tts_use_guide_tokens:
-        type: boolean
-      typical:
-        type: number
-      ubatch_size:
-        type: integer
-      verbose:
-        type: boolean
-      verbose_prompt:
-        description: Common params
-        type: boolean
-      verbosity:
-        type: integer
-      xtc_probability:
-        type: number
-      xtc_threshold:
-        type: number
-      yarn_attn_factor:
-        type: number
-      yarn_beta_fast:
-        type: number
-      yarn_beta_slow:
-        type: number
-      yarn_ext_factor:
-        type: number
-      yarn_orig_ctx:
+      restart_delay:
+        description: seconds
         type: integer
     type: object
+  instance.InstanceStatus:
+    enum:
+    - 0
+    - 1
+    - 2
+    type: integer
+    x-enum-varnames:
+    - Stopped
+    - Running
+    - Failed
   instance.Process:
     properties:
       created:
@@ -351,9 +49,10 @@ definitions:
         type: integer
       name:
         type: string
-      running:
+      status:
+        allOf:
+        - $ref: '#/definitions/instance.InstanceStatus'
         description: Status
-        type: boolean
     type: object
   server.OpenAIInstance:
     properties:
@@ -375,6 +74,11 @@ definitions:
       object:
         type: string
     type: object
+  server.ParseCommandRequest:
+    properties:
+      command:
+        type: string
+    type: object
 info:
   contact: {}
   description: llamactl is a control server for managing Llama Server instances.
@@ -384,6 +88,153 @@ info:
   title: llamactl API
   version: "1.0"
 paths:
+  /backends/llama-cpp/devices:
+    get:
+      description: Returns a list of available devices for the llama server
+      responses:
+        "200":
+          description: List of devices
+          schema:
+            type: string
+        "500":
+          description: Internal Server Error
+          schema:
+            type: string
+      security:
+      - ApiKeyAuth: []
+      summary: List available devices for llama server
+      tags:
+      - backends
+  /backends/llama-cpp/help:
+    get:
+      description: Returns the help text for the llama server command
+      responses:
+        "200":
+          description: Help text
+          schema:
+            type: string
+        "500":
+          description: Internal Server Error
+          schema:
+            type: string
+      security:
+      - ApiKeyAuth: []
+      summary: Get help for llama server
+      tags:
+      - backends
+  /backends/llama-cpp/parse-command:
+    post:
+      consumes:
+      - application/json
+      description: Parses a llama-server command string into instance options
+      parameters:
+      - description: Command to parse
+        in: body
+        name: request
+        required: true
+        schema:
+          $ref: '#/definitions/server.ParseCommandRequest'
+      produces:
+      - application/json
+      responses:
+        "200":
+          description: Parsed options
+          schema:
+            $ref: '#/definitions/instance.CreateInstanceOptions'
+        "400":
+          description: Invalid request or command
+          schema:
+            additionalProperties:
+              type: string
+            type: object
+        "500":
+          description: Internal Server Error
+          schema:
+            additionalProperties:
+              type: string
+            type: object
+      security:
+      - ApiKeyAuth: []
+      summary: Parse llama-server command
+      tags:
+      - backends
+  /backends/llama-cpp/version:
+    get:
+      description: Returns the version of the llama server command
+      responses:
+        "200":
+          description: Version information
+          schema:
+            type: string
+        "500":
+          description: Internal Server Error
+          schema:
+            type: string
+      security:
+      - ApiKeyAuth: []
+      summary: Get version of llama server
+      tags:
+      - backends
+  /backends/mlx/parse-command:
+    post:
+      consumes:
+      - application/json
+      description: Parses MLX-LM server command string into instance options
+      parameters:
+      - description: Command to parse
+        in: body
+        name: request
+        required: true
+        schema:
+          $ref: '#/definitions/server.ParseCommandRequest'
+      produces:
+      - application/json
+      responses:
+        "200":
+          description: Parsed options
+          schema:
+            $ref: '#/definitions/instance.CreateInstanceOptions'
+        "400":
+          description: Invalid request or command
+          schema:
+            additionalProperties:
+              type: string
+            type: object
+      security:
+      - ApiKeyAuth: []
+      summary: Parse mlx_lm.server command
+      tags:
+      - backends
+  /backends/vllm/parse-command:
+    post:
+      consumes:
+      - application/json
+      description: Parses a vLLM serve command string into instance options
+      parameters:
+      - description: Command to parse
+        in: body
+        name: request
+        required: true
+        schema:
+          $ref: '#/definitions/server.ParseCommandRequest'
+      produces:
+      - application/json
+      responses:
+        "200":
+          description: Parsed options
+          schema:
+            $ref: '#/definitions/instance.CreateInstanceOptions'
+        "400":
+          description: Invalid request or command
+          schema:
+            additionalProperties:
+              type: string
+            type: object
+      security:
+      - ApiKeyAuth: []
+      summary: Parse vllm serve command
+      tags:
+      - backends
   /instances:
     get:
       description: Returns a list of all instances managed by the server
@@ -694,57 +545,6 @@ paths:
       summary: Stop a running instance
       tags:
       - instances
-  /server/devices:
-    get:
-      description: Returns a list of available devices for the llama server
-      responses:
-        "200":
-          description: List of devices
-          schema:
-            type: string
-        "500":
-          description: Internal Server Error
-          schema:
-            type: string
-      security:
-      - ApiKeyAuth: []
-      summary: List available devices for llama server
-      tags:
-      - server
-  /server/help:
-    get:
-      description: Returns the help text for the llama server command
-      responses:
-        "200":
-          description: Help text
-          schema:
-            type: string
-        "500":
-          description: Internal Server Error
-          schema:
-            type: string
-      security:
-      - ApiKeyAuth: []
-      summary: Get help for llama server
-      tags:
-      - server
-  /server/version:
-    get:
-      description: Returns the version of the llama server command
-      responses:
-        "200":
-          description: Version information
-          schema:
-            type: string
-        "500":
-          description: Internal Server Error
-          schema:
-            type: string
-      security:
-      - ApiKeyAuth: []
-      summary: Get version of llama server
-      tags:
-      - server
   /v1/:
     post:
       consumes:
@@ -756,7 +556,7 @@ paths:
         "200":
           description: OpenAI response
         "400":
-          description: Invalid request body or model name
+          description: Invalid request body or instance name
           schema:
             type: string
         "500":

cmd/server/main.go

@@ -58,7 +58,7 @@ func main() {
 	}
 
 	// Initialize the instance manager
-	instanceManager := manager.NewInstanceManager(cfg.Instances)
+	instanceManager := manager.NewInstanceManager(cfg.Backends, cfg.Instances)
 
 	// Create a new handler with the instance manager
 	handler := server.NewHandler(instanceManager, cfg)

docker/Dockerfile.llamacpp

@@ -0,0 +1,23 @@
+FROM ghcr.io/ggml-org/llama.cpp:server-cuda
+
+# Install curl for downloading llamactl
+RUN apt-get update && apt-get install -y curl && rm -rf /var/lib/apt/lists/*
+
+# Download and install the latest llamactl release
+RUN LATEST_VERSION=$(curl -s https://api.github.com/repos/lordmathis/llamactl/releases/latest | grep '"tag_name":' | sed -E 's/.*"([^"]+)".*/\1/') && \
+    curl -L "https://github.com/lordmathis/llamactl/releases/download/${LATEST_VERSION}/llamactl-${LATEST_VERSION}-linux-amd64.tar.gz" | tar -xz && \
+    mv llamactl /usr/local/bin/ && \
+    chmod +x /usr/local/bin/llamactl
+
+# Set working directory
+RUN mkdir -p /data
+WORKDIR /data
+
+# Expose the default llamactl port
+EXPOSE 8080
+
+ENV LLAMACTL_LLAMACPP_COMMAND=/app/llama-server
+ENV LD_LIBRARY_PATH="/app:/usr/local/lib:/usr/lib"
+
+# Set llamactl as the entrypoint
+ENTRYPOINT ["llamactl"]

docker/Dockerfile.source

@@ -0,0 +1,64 @@
+# WebUI build stage
+FROM node:20-alpine AS webui-builder
+
+WORKDIR /webui
+
+# Copy webui package files
+COPY webui/package*.json ./
+
+# Install dependencies
+RUN npm ci
+
+# Copy webui source
+COPY webui/ ./
+
+# Build webui
+RUN npm run build
+
+# Go build stage
+FROM golang:1.24-alpine AS builder
+
+# Install build dependencies
+RUN apk add --no-cache git ca-certificates
+
+# Set working directory
+WORKDIR /build
+
+# Copy go mod files
+COPY go.mod go.sum ./
+
+# Download dependencies
+RUN go mod download
+
+# Copy source code
+COPY cmd/ ./cmd/
+COPY pkg/ ./pkg/
+COPY apidocs/ ./apidocs/
+COPY webui/webui.go ./webui/
+
+# Copy built webui from webui-builder
+COPY --from=webui-builder /webui/dist ./webui/dist
+
+# Build the application
+RUN CGO_ENABLED=0 GOOS=linux go build -a -installsuffix cgo -ldflags="-w -s" -o llamactl ./cmd/server
+
+# Final stage
+FROM alpine:latest
+
+# Install runtime dependencies
+RUN apk --no-cache add ca-certificates
+
+# Create data directory
+RUN mkdir -p /data
+
+# Set working directory
+WORKDIR /data
+
+# Copy binary from builder
+COPY --from=builder /build/llamactl /usr/local/bin/llamactl
+
+# Expose the default port
+EXPOSE 8080
+
+# Set llamactl as the entrypoint
+ENTRYPOINT ["llamactl"]

docker/Dockerfile.vllm

@@ -0,0 +1,20 @@
+FROM vllm/vllm-openai:latest
+
+# Install curl for downloading llamactl
+RUN apt-get update && apt-get install -y curl && rm -rf /var/lib/apt/lists/*
+
+# Download and install the latest llamactl release
+RUN LATEST_VERSION=$(curl -s https://api.github.com/repos/lordmathis/llamactl/releases/latest | grep '"tag_name":' | sed -E 's/.*"([^"]+)".*/\1/') && \
+    curl -L "https://github.com/lordmathis/llamactl/releases/download/${LATEST_VERSION}/llamactl-${LATEST_VERSION}-linux-amd64.tar.gz" | tar -xz && \
+    mv llamactl /usr/local/bin/ && \
+    chmod +x /usr/local/bin/llamactl
+
+# Set working directory
+RUN mkdir -p /data
+WORKDIR /data
+
+# Expose the default llamactl port
+EXPOSE 8080
+
+# Set llamactl as the entrypoint
+ENTRYPOINT ["llamactl"]

docker/docker-compose.yml

@@ -0,0 +1,56 @@
+version: '3.8'
+
+services:
+  llamactl-llamacpp:
+    build:
+      context: ..
+      dockerfile: docker/Dockerfile.llamacpp
+    image: llamactl:llamacpp-cuda
+    container_name: llamactl-llamacpp
+    ports:
+      - "8080:8080"
+    volumes:
+      - ./data/llamacpp:/data
+      - ./models:/models  # Mount models directory
+      - ~/.cache/llama.cpp:/root/.cache/llama.cpp  # Llama.cpp cache
+    environment:
+      # Set data directory for persistence
+      - LLAMACTL_DATA_DIR=/data
+      # Enable Docker mode for nested containers (if needed)
+      - LLAMACTL_LLAMACPP_DOCKER_ENABLED=false
+    deploy:
+      resources:
+        reservations:
+          devices:
+            - driver: nvidia
+              count: all
+              capabilities: [gpu]
+    restart: unless-stopped
+
+  llamactl-vllm:
+    build:
+      context: ..
+      dockerfile: docker/Dockerfile.vllm
+    image: llamactl:vllm-cuda
+    container_name: llamactl-vllm
+    ports:
+      - "8081:8080"  # Use different port to avoid conflicts
+    volumes:
+      - ./data/vllm:/data
+      - ./models:/models  # Mount models directory
+      - ~/.cache/huggingface:/root/.cache/huggingface  # HuggingFace cache
+    environment:
+      # Set data directory for persistence
+      - LLAMACTL_DATA_DIR=/data
+      # Enable Docker mode for nested containers (if needed)
+      - LLAMACTL_VLLM_DOCKER_ENABLED=false
+      # vLLM specific environment variables
+      - CUDA_VISIBLE_DEVICES=all
+    deploy:
+      resources:
+        reservations:
+          devices:
+            - driver: nvidia
+              count: all
+              capabilities: [gpu]
+    restart: unless-stopped

docs-requirements.txt

@@ -0,0 +1,5 @@
+mkdocs-material==9.5.3
+mkdocs==1.5.3
+pymdown-extensions==10.7
+mkdocs-git-revision-date-localized-plugin==1.2.4
+mike==2.0.0

docs/getting-started/configuration.md

@@ -0,0 +1,252 @@
+# Configuration
+
+llamactl can be configured via configuration files or environment variables. Configuration is loaded in the following order of precedence:
+
+```
+Defaults < Configuration file < Environment variables
+```
+
+llamactl works out of the box with sensible defaults, but you can customize the behavior to suit your needs.
+
+## Default Configuration
+
+Here's the default configuration with all available options:
+
+```yaml
+server:
+  host: "0.0.0.0"                # Server host to bind to
+  port: 8080                     # Server port to bind to
+  allowed_origins: ["*"]         # Allowed CORS origins (default: all)
+  allowed_headers: ["*"]         # Allowed CORS headers (default: all)
+  enable_swagger: false          # Enable Swagger UI for API docs
+
+backends:
+  llama-cpp:
+    command: "llama-server"
+    args: []
+    environment: {}              # Environment variables for the backend process
+    docker:
+      enabled: false
+      image: "ghcr.io/ggml-org/llama.cpp:server"
+      args: ["run", "--rm", "--network", "host", "--gpus", "all"]
+      environment: {}
+    response_headers: {}         # Additional response headers to send with responses
+
+  vllm:
+    command: "vllm"
+    args: ["serve"]
+    environment: {}              # Environment variables for the backend process
+    docker:
+      enabled: false
+      image: "vllm/vllm-openai:latest"
+      args: ["run", "--rm", "--network", "host", "--gpus", "all", "--shm-size", "1g"]
+      environment: {}
+    response_headers: {}         # Additional response headers to send with responses
+
+  mlx:
+    command: "mlx_lm.server"
+    args: []
+    environment: {}              # Environment variables for the backend process
+    response_headers: {}         # Additional response headers to send with responses
+
+instances:
+  port_range: [8000, 9000]       # Port range for instances
+  data_dir: ~/.local/share/llamactl         # Data directory (platform-specific, see below)
+  configs_dir: ~/.local/share/llamactl/instances  # Instance configs directory
+  logs_dir: ~/.local/share/llamactl/logs    # Logs directory
+  auto_create_dirs: true         # Auto-create data/config/logs dirs if missing
+  max_instances: -1              # Max instances (-1 = unlimited)
+  max_running_instances: -1      # Max running instances (-1 = unlimited)
+  enable_lru_eviction: true      # Enable LRU eviction for idle instances
+  default_auto_restart: true     # Auto-restart new instances by default
+  default_max_restarts: 3        # Max restarts for new instances
+  default_restart_delay: 5       # Restart delay (seconds) for new instances
+  default_on_demand_start: true  # Default on-demand start setting
+  on_demand_start_timeout: 120   # Default on-demand start timeout in seconds
+  timeout_check_interval: 5      # Idle instance timeout check in minutes
+
+auth:
+  require_inference_auth: true   # Require auth for inference endpoints
+  inference_keys: []             # Keys for inference endpoints
+  require_management_auth: true  # Require auth for management endpoints
+  management_keys: []            # Keys for management endpoints
+```
+
+## Configuration Files
+
+### Configuration File Locations
+
+Configuration files are searched in the following locations (in order of precedence):
+
+**Linux:**  
+- `./llamactl.yaml` or `./config.yaml` (current directory)  
+- `$HOME/.config/llamactl/config.yaml`  
+- `/etc/llamactl/config.yaml`  
+
+**macOS:**  
+- `./llamactl.yaml` or `./config.yaml` (current directory)  
+- `$HOME/Library/Application Support/llamactl/config.yaml`  
+- `/Library/Application Support/llamactl/config.yaml`  
+
+**Windows:**  
+- `./llamactl.yaml` or `./config.yaml` (current directory)  
+- `%APPDATA%\llamactl\config.yaml`  
+- `%USERPROFILE%\llamactl\config.yaml`  
+- `%PROGRAMDATA%\llamactl\config.yaml`  
+
+You can specify the path to config file with `LLAMACTL_CONFIG_PATH` environment variable.
+
+## Configuration Options
+
+### Server Configuration
+
+```yaml
+server:
+  host: "0.0.0.0"         # Server host to bind to (default: "0.0.0.0")
+  port: 8080              # Server port to bind to (default: 8080)
+  allowed_origins: ["*"]  # CORS allowed origins (default: ["*"])
+  allowed_headers: ["*"]  # CORS allowed headers (default: ["*"])
+  enable_swagger: false   # Enable Swagger UI (default: false)
+```
+
+**Environment Variables:**
+- `LLAMACTL_HOST` - Server host
+- `LLAMACTL_PORT` - Server port
+- `LLAMACTL_ALLOWED_ORIGINS` - Comma-separated CORS origins
+- `LLAMACTL_ENABLE_SWAGGER` - Enable Swagger UI (true/false)
+
+### Backend Configuration
+```yaml
+backends:
+  llama-cpp:
+    command: "llama-server"
+    args: []
+    environment: {}              # Environment variables for the backend process
+    docker:
+      enabled: false             # Enable Docker runtime (default: false)
+      image: "ghcr.io/ggml-org/llama.cpp:server"
+      args: ["run", "--rm", "--network", "host", "--gpus", "all"]
+      environment: {}
+    response_headers: {}         # Additional response headers to send with responses
+
+  vllm:
+    command: "vllm"
+    args: ["serve"]
+    environment: {}              # Environment variables for the backend process
+    docker:
+      enabled: false             # Enable Docker runtime (default: false)
+      image: "vllm/vllm-openai:latest"
+      args: ["run", "--rm", "--network", "host", "--gpus", "all", "--shm-size", "1g"]
+      environment: {}
+    response_headers: {}         # Additional response headers to send with responses
+
+  mlx:
+    command: "mlx_lm.server"
+    args: []
+    environment: {}              # Environment variables for the backend process
+    # MLX does not support Docker
+    response_headers: {}         # Additional response headers to send with responses
+```
+
+**Backend Configuration Fields:**
+- `command`: Executable name/path for the backend
+- `args`: Default arguments prepended to all instances
+- `environment`: Environment variables for the backend process (optional)
+- `response_headers`: Additional response headers to send with responses (optional)
+- `docker`: Docker-specific configuration (optional)
+  - `enabled`: Boolean flag to enable Docker runtime
+  - `image`: Docker image to use
+  - `args`: Additional arguments passed to `docker run`
+  - `environment`: Environment variables for the container (optional)
+
+> If llamactl is behind an NGINX proxy, `X-Accel-Buffering: no` response header may be required for NGINX to properly stream the responses without buffering.
+
+**Environment Variables:**
+
+**LlamaCpp Backend:**
+- `LLAMACTL_LLAMACPP_COMMAND` - LlamaCpp executable command
+- `LLAMACTL_LLAMACPP_ARGS` - Space-separated default arguments
+- `LLAMACTL_LLAMACPP_ENV` - Environment variables in format "KEY1=value1,KEY2=value2"
+- `LLAMACTL_LLAMACPP_DOCKER_ENABLED` - Enable Docker runtime (true/false)
+- `LLAMACTL_LLAMACPP_DOCKER_IMAGE` - Docker image to use
+- `LLAMACTL_LLAMACPP_DOCKER_ARGS` - Space-separated Docker arguments
+- `LLAMACTL_LLAMACPP_DOCKER_ENV` - Docker environment variables in format "KEY1=value1,KEY2=value2"
+- `LLAMACTL_LLAMACPP_RESPONSE_HEADERS` - Response headers in format "KEY1=value1;KEY2=value2"
+
+**VLLM Backend:**
+- `LLAMACTL_VLLM_COMMAND` - VLLM executable command
+- `LLAMACTL_VLLM_ARGS` - Space-separated default arguments
+- `LLAMACTL_VLLM_ENV` - Environment variables in format "KEY1=value1,KEY2=value2"
+- `LLAMACTL_VLLM_DOCKER_ENABLED` - Enable Docker runtime (true/false)
+- `LLAMACTL_VLLM_DOCKER_IMAGE` - Docker image to use
+- `LLAMACTL_VLLM_DOCKER_ARGS` - Space-separated Docker arguments
+- `LLAMACTL_VLLM_DOCKER_ENV` - Docker environment variables in format "KEY1=value1,KEY2=value2"
+- `LLAMACTL_VLLM_RESPONSE_HEADERS` - Response headers in format "KEY1=value1;KEY2=value2"
+
+**MLX Backend:**
+- `LLAMACTL_MLX_COMMAND` - MLX executable command
+- `LLAMACTL_MLX_ARGS` - Space-separated default arguments
+- `LLAMACTL_MLX_ENV` - Environment variables in format "KEY1=value1,KEY2=value2"
+- `LLAMACTL_MLX_RESPONSE_HEADERS` - Response headers in format "KEY1=value1;KEY2=value2"
+
+### Instance Configuration
+
+```yaml
+instances:
+  port_range: [8000, 9000]                          # Port range for instances (default: [8000, 9000])
+  data_dir: "~/.local/share/llamactl"               # Directory for all llamactl data (default varies by OS)
+  configs_dir: "~/.local/share/llamactl/instances"  # Directory for instance configs (default: data_dir/instances)
+  logs_dir: "~/.local/share/llamactl/logs"          # Directory for instance logs (default: data_dir/logs)
+  auto_create_dirs: true                            # Automatically create data/config/logs directories (default: true)
+  max_instances: -1                                 # Maximum instances (-1 = unlimited)
+  max_running_instances: -1                         # Maximum running instances (-1 = unlimited)
+  enable_lru_eviction: true                         # Enable LRU eviction for idle instances
+  default_auto_restart: true                        # Default auto-restart setting
+  default_max_restarts: 3                           # Default maximum restart attempts
+  default_restart_delay: 5                          # Default restart delay in seconds
+  default_on_demand_start: true                     # Default on-demand start setting
+  on_demand_start_timeout: 120                      # Default on-demand start timeout in seconds
+  timeout_check_interval: 5                         # Default instance timeout check interval in minutes
+```
+
+**Environment Variables:**  
+- `LLAMACTL_INSTANCE_PORT_RANGE` - Port range (format: "8000-9000" or "8000,9000")  
+- `LLAMACTL_DATA_DIRECTORY` - Data directory path  
+- `LLAMACTL_INSTANCES_DIR` - Instance configs directory path  
+- `LLAMACTL_LOGS_DIR` - Log directory path  
+- `LLAMACTL_AUTO_CREATE_DATA_DIR` - Auto-create data/config/logs directories (true/false)  
+- `LLAMACTL_MAX_INSTANCES` - Maximum number of instances  
+- `LLAMACTL_MAX_RUNNING_INSTANCES` - Maximum number of running instances
+- `LLAMACTL_ENABLE_LRU_EVICTION` - Enable LRU eviction for idle instances
+- `LLAMACTL_DEFAULT_AUTO_RESTART` - Default auto-restart setting (true/false)  
+- `LLAMACTL_DEFAULT_MAX_RESTARTS` - Default maximum restarts  
+- `LLAMACTL_DEFAULT_RESTART_DELAY` - Default restart delay in seconds  
+- `LLAMACTL_DEFAULT_ON_DEMAND_START` - Default on-demand start setting (true/false)  
+- `LLAMACTL_ON_DEMAND_START_TIMEOUT` - Default on-demand start timeout in seconds  
+- `LLAMACTL_TIMEOUT_CHECK_INTERVAL` - Default instance timeout check interval in minutes  
+
+### Authentication Configuration
+
+```yaml
+auth:
+  require_inference_auth: true           # Require API key for OpenAI endpoints (default: true)
+  inference_keys: []                     # List of valid inference API keys
+  require_management_auth: true          # Require API key for management endpoints (default: true)
+  management_keys: []                    # List of valid management API keys
+```
+
+**Environment Variables:**  
+- `LLAMACTL_REQUIRE_INFERENCE_AUTH` - Require auth for OpenAI endpoints (true/false)  
+- `LLAMACTL_INFERENCE_KEYS` - Comma-separated inference API keys  
+- `LLAMACTL_REQUIRE_MANAGEMENT_AUTH` - Require auth for management endpoints (true/false)  
+- `LLAMACTL_MANAGEMENT_KEYS` - Comma-separated management API keys  
+
+## Command Line Options
+
+View all available command line options:
+
+```bash
+llamactl --help
+```
+
+You can also override configuration using command line flags when starting llamactl.

docs/getting-started/installation.md

@@ -0,0 +1,170 @@
+# Installation
+
+This guide will walk you through installing Llamactl on your system.
+
+## Prerequisites
+
+### Backend Dependencies
+
+llamactl supports multiple backends. Install at least one:
+
+**For llama.cpp backend (all platforms):**
+
+You need `llama-server` from [llama.cpp](https://github.com/ggml-org/llama.cpp) installed:
+
+```bash
+# Homebrew (macOS/Linux)
+brew install llama.cpp
+# Winget (Windows)
+winget install llama.cpp
+```
+
+Or build from source - see llama.cpp docs
+
+**For MLX backend (macOS only):**
+
+MLX provides optimized inference on Apple Silicon. Install MLX-LM:
+
+```bash
+# Install via pip (requires Python 3.8+)
+pip install mlx-lm
+
+# Or in a virtual environment (recommended)
+python -m venv mlx-env
+source mlx-env/bin/activate
+pip install mlx-lm
+```
+
+Note: MLX backend is only available on macOS with Apple Silicon (M1, M2, M3, etc.)
+
+**For vLLM backend:**
+
+vLLM provides high-throughput distributed serving for LLMs. Install vLLM:
+
+```bash
+# Install via pip (requires Python 3.8+, GPU required)
+pip install vllm
+
+# Or in a virtual environment (recommended)
+python -m venv vllm-env
+source vllm-env/bin/activate
+pip install vllm
+
+# For production deployments, consider container-based installation
+```
+
+## Installation Methods
+
+### Option 1: Download Binary (Recommended)
+
+Download the latest release from the [GitHub releases page](https://github.com/lordmathis/llamactl/releases):
+
+```bash
+# Linux/macOS - Get latest version and download
+LATEST_VERSION=$(curl -s https://api.github.com/repos/lordmathis/llamactl/releases/latest | grep '"tag_name":' | sed -E 's/.*"([^"]+)".*/\1/')
+curl -L https://github.com/lordmathis/llamactl/releases/download/${LATEST_VERSION}/llamactl-${LATEST_VERSION}-$(uname -s | tr '[:upper:]' '[:lower:]')-$(uname -m).tar.gz | tar -xz
+sudo mv llamactl /usr/local/bin/
+
+# Or download manually from:
+# https://github.com/lordmathis/llamactl/releases/latest
+
+# Windows - Download from releases page
+```
+
+### Option 2: Docker
+
+llamactl provides Dockerfiles for creating Docker images with backends pre-installed. The resulting images include the latest llamactl release with the respective backend.
+
+**Available Dockerfiles (CUDA):**
+- **llamactl with llama.cpp CUDA**: `docker/Dockerfile.llamacpp` (based on `ghcr.io/ggml-org/llama.cpp:server-cuda`)
+- **llamactl with vLLM CUDA**: `docker/Dockerfile.vllm` (based on `vllm/vllm-openai:latest`)
+- **llamactl built from source**: `docker/Dockerfile.source` (multi-stage build with webui)
+
+**Note:** These Dockerfiles are configured for CUDA. For other platforms (CPU, ROCm, Vulkan, etc.), adapt the base image. For llama.cpp, see available tags at [llama.cpp Docker docs](https://github.com/ggml-org/llama.cpp/blob/master/docs/docker.md). For vLLM, check [vLLM docs](https://docs.vllm.ai/en/v0.6.5/serving/deploying_with_docker.html).
+
+#### Using Docker Compose
+
+```bash
+# Clone the repository
+git clone https://github.com/lordmathis/llamactl.git
+cd llamactl
+
+# Create directories for data and models
+mkdir -p data/llamacpp data/vllm models
+
+# Start llamactl with llama.cpp backend
+docker-compose -f docker/docker-compose.yml up llamactl-llamacpp -d
+
+# Or start llamactl with vLLM backend
+docker-compose -f docker/docker-compose.yml up llamactl-vllm -d
+```
+
+Access the dashboard at:
+- llamactl with llama.cpp: http://localhost:8080
+- llamactl with vLLM: http://localhost:8081
+
+#### Using Docker Build and Run
+
+**llamactl with llama.cpp CUDA:**
+```bash
+docker build -f docker/Dockerfile.llamacpp -t llamactl:llamacpp-cuda .
+docker run -d \
+  --name llamactl-llamacpp \
+  --gpus all \
+  -p 8080:8080 \
+  -v ~/.cache/llama.cpp:/root/.cache/llama.cpp \
+  llamactl:llamacpp-cuda
+```
+
+**llamactl with vLLM CUDA:**
+```bash
+docker build -f docker/Dockerfile.vllm -t llamactl:vllm-cuda .
+docker run -d \
+  --name llamactl-vllm \
+  --gpus all \
+  -p 8080:8080 \
+  -v ~/.cache/huggingface:/root/.cache/huggingface \
+  llamactl:vllm-cuda
+```
+
+**llamactl built from source:**
+```bash
+docker build -f docker/Dockerfile.source -t llamactl:source .
+docker run -d \
+  --name llamactl \
+  -p 8080:8080 \
+  llamactl:source
+```
+
+### Option 3: Build from Source
+
+Requirements:
+- Go 1.24 or later
+- Node.js 22 or later
+- Git
+
+If you prefer to build from source:
+
+```bash
+# Clone the repository
+git clone https://github.com/lordmathis/llamactl.git
+cd llamactl
+
+# Build the web UI
+cd webui && npm ci && npm run build && cd ..
+
+# Build the application
+go build -o llamactl ./cmd/server
+```
+
+## Verification
+
+Verify your installation by checking the version:
+
+```bash
+llamactl --version
+```
+
+## Next Steps
+
+Now that Llamactl is installed, continue to the [Quick Start](quick-start.md) guide to get your first instance running!

docs/getting-started/quick-start.md

@@ -0,0 +1,190 @@
+# Quick Start
+
+This guide will help you get Llamactl up and running in just a few minutes.
+
+## Step 1: Start Llamactl
+
+Start the Llamactl server:
+
+```bash
+llamactl
+```
+
+By default, Llamactl will start on `http://localhost:8080`.
+
+## Step 2: Access the Web UI
+
+Open your web browser and navigate to:
+
+```
+http://localhost:8080
+```
+
+Login with the management API key. By default it is generated during server startup. Copy it from the terminal output.
+
+You should see the Llamactl web interface.
+
+## Step 3: Create Your First Instance
+
+1. Click the "Add Instance" button
+2. Fill in the instance configuration:
+   - **Name**: Give your instance a descriptive name
+   - **Backend Type**: Choose from llama.cpp, MLX, or vLLM
+   - **Model**: Model path or identifier for your chosen backend
+   - **Additional Options**: Backend-specific parameters
+
+3. Click "Create Instance"
+
+## Step 4: Start Your Instance
+
+Once created, you can:
+
+- **Start** the instance by clicking the start button
+- **Monitor** its status in real-time
+- **View logs** by clicking the logs button
+- **Stop** the instance when needed
+
+## Example Configurations
+
+Here are basic example configurations for each backend:
+
+**llama.cpp backend:**
+```json
+{
+  "name": "llama2-7b",
+  "backend_type": "llama_cpp",
+  "backend_options": {
+    "model": "/path/to/llama-2-7b-chat.gguf",
+    "threads": 4,
+    "ctx_size": 2048,
+    "gpu_layers": 32
+  }
+}
+```
+
+**MLX backend (macOS only):**
+```json
+{
+  "name": "mistral-mlx",
+  "backend_type": "mlx_lm",
+  "backend_options": {
+    "model": "mlx-community/Mistral-7B-Instruct-v0.3-4bit",
+    "temp": 0.7,
+    "max_tokens": 2048
+  }
+}
+```
+
+**vLLM backend:**
+```json
+{
+  "name": "dialogpt-vllm",
+  "backend_type": "vllm",
+  "backend_options": {
+    "model": "microsoft/DialoGPT-medium",
+    "tensor_parallel_size": 2,
+    "gpu_memory_utilization": 0.9
+  }
+}
+```
+
+## Docker Support
+
+Llamactl can run backends in Docker containers. To enable Docker for a backend, add a `docker` section to that backend in your YAML configuration file (e.g. `config.yaml`) as shown below:
+
+```yaml
+backends:
+  vllm:
+    command: "vllm"
+    args: ["serve"]
+    docker:
+      enabled: true
+      image: "vllm/vllm-openai:latest"
+      args: ["run", "--rm", "--network", "host", "--gpus", "all", "--shm-size", "1g"]
+```
+
+## Using the API
+
+You can also manage instances via the REST API:
+
+```bash
+# List all instances
+curl http://localhost:8080/api/instances
+
+# Create a new llama.cpp instance
+curl -X POST http://localhost:8080/api/instances/my-model \
+  -H "Content-Type: application/json" \
+  -d '{
+    "backend_type": "llama_cpp",
+    "backend_options": {
+      "model": "/path/to/model.gguf"
+    }
+  }'
+
+# Start an instance
+curl -X POST http://localhost:8080/api/instances/my-model/start
+```
+
+## OpenAI Compatible API
+
+Llamactl provides OpenAI-compatible endpoints, making it easy to integrate with existing OpenAI client libraries and tools.
+
+### Chat Completions
+
+Once you have an instance running, you can use it with the OpenAI-compatible chat completions endpoint:
+
+```bash
+curl -X POST http://localhost:8080/v1/chat/completions \
+  -H "Content-Type: application/json" \
+  -d '{
+    "model": "my-model",
+    "messages": [
+      {
+        "role": "user",
+        "content": "Hello! Can you help me write a Python function?"
+      }
+    ],
+    "max_tokens": 150,
+    "temperature": 0.7
+  }'
+```
+
+### Using with Python OpenAI Client
+
+You can also use the official OpenAI Python client:
+
+```python
+from openai import OpenAI
+
+# Point the client to your Llamactl server
+client = OpenAI(
+    base_url="http://localhost:8080/v1",
+    api_key="not-needed"  # Llamactl doesn't require API keys by default
+)
+
+# Create a chat completion
+response = client.chat.completions.create(
+    model="my-model",  # Use the name of your instance
+    messages=[
+        {"role": "user", "content": "Explain quantum computing in simple terms"}
+    ],
+    max_tokens=200,
+    temperature=0.7
+)
+
+print(response.choices[0].message.content)
+```
+
+### List Available Models
+
+Get a list of running instances (models) in OpenAI-compatible format:
+
+```bash
+curl http://localhost:8080/v1/models
+```
+
+## Next Steps
+
+- Manage instances [Managing Instances](../user-guide/managing-instances.md)
+- Explore the [API Reference](../user-guide/api-reference.md)
+- Configure advanced settings in the [Configuration](configuration.md) guide

docs/index.md

@@ -0,0 +1,34 @@
+# Llamactl Documentation
+
+Welcome to the Llamactl documentation!  
+
+![Dashboard Screenshot](images/dashboard.png)
+
+## What is Llamactl?
+
+**{{HEADLINE}}**
+
+## Features
+
+{{FEATURES}}  
+
+## Quick Links
+
+- [Installation Guide](getting-started/installation.md) - Get Llamactl up and running
+- [Configuration Guide](getting-started/configuration.md) - Detailed configuration options
+- [Quick Start](getting-started/quick-start.md) - Your first steps with Llamactl
+- [Managing Instances](user-guide/managing-instances.md) - Instance lifecycle management
+- [API Reference](user-guide/api-reference.md) - Complete API documentation
+
+
+## Getting Help
+
+If you need help or have questions:
+
+- Check the [Troubleshooting](user-guide/troubleshooting.md) guide
+- Visit the [GitHub repository](https://github.com/lordmathis/llamactl)
+- Review the [Configuration Guide](getting-started/configuration.md) for advanced settings
+
+## License
+
+MIT License - see the [LICENSE](https://github.com/lordmathis/llamactl/blob/main/LICENSE) file.

docs/readme_sync.py

@@ -0,0 +1,62 @@
+"""
+MkDocs hook to sync content from README.md to docs/index.md
+"""
+import re
+import os
+
+
+def on_page_markdown(markdown, page, config, **kwargs):
+    """Process markdown content before rendering"""
+    # Only process the index.md file
+    if page.file.src_path != 'index.md':
+        return markdown
+
+    # Get the path to README.md (relative to mkdocs.yml)
+    readme_path = os.path.join(os.path.dirname(config['config_file_path']), 'README.md')
+
+    if not os.path.exists(readme_path):
+        print(f"Warning: README.md not found at {readme_path}")
+        return markdown
+
+    try:
+        with open(readme_path, 'r', encoding='utf-8') as f:
+            readme_content = f.read()
+    except Exception as e:
+        print(f"Error reading README.md: {e}")
+        return markdown
+
+    # Extract headline (the text in bold after the title)
+    headline_match = re.search(r'\*\*(.*?)\*\*', readme_content)
+    headline = headline_match.group(1) if headline_match else 'Management server for llama.cpp and MLX instances'
+
+    # Extract features section - everything between ## Features and the next ## heading
+    features_match = re.search(r'## Features\n(.*?)(?=\n## |\Z)', readme_content, re.DOTALL)
+    if features_match:
+        features_content = features_match.group(1).strip()
+        # Just add line breaks at the end of each line for proper MkDocs rendering
+        features_with_breaks = add_line_breaks(features_content)
+    else:
+        features_with_breaks = "Features content not found in README.md"
+
+    # Replace placeholders in the markdown
+    markdown = markdown.replace('{{HEADLINE}}', headline)
+    markdown = markdown.replace('{{FEATURES}}', features_with_breaks)
+
+    # Fix image paths: convert docs/images/ to images/ for MkDocs
+    markdown = re.sub(r'docs/images/', 'images/', markdown)
+
+    return markdown
+
+
+def add_line_breaks(content):
+    """Add two spaces at the end of each line for proper MkDocs line breaks"""
+    lines = content.split('\n')
+    processed_lines = []
+
+    for line in lines:
+        if line.strip():  # Only add spaces to non-empty lines
+            processed_lines.append(line.rstrip() + '  ')
+        else:
+            processed_lines.append(line)
+
+    return '\n'.join(processed_lines)

docs/user-guide/api-reference.md

@@ -0,0 +1,527 @@
+# API Reference
+
+Complete reference for the Llamactl REST API.
+
+## Base URL
+
+All API endpoints are relative to the base URL:
+
+```
+http://localhost:8080/api/v1
+```
+
+## Authentication
+
+Llamactl supports API key authentication. If authentication is enabled, include the API key in the Authorization header:
+
+```bash
+curl -H "Authorization: Bearer <your-api-key>" \
+  http://localhost:8080/api/v1/instances
+```
+
+The server supports two types of API keys:
+- **Management API Keys**: Required for instance management operations (CRUD operations on instances)
+- **Inference API Keys**: Required for OpenAI-compatible inference endpoints
+
+## System Endpoints
+
+### Get Llamactl Version
+
+Get the version information of the llamactl server.
+
+```http
+GET /api/v1/version
+```
+
+**Response:**
+```
+Version: 1.0.0
+Commit: abc123
+Build Time: 2024-01-15T10:00:00Z
+```
+
+### Get Llama Server Help
+
+Get help text for the llama-server command.
+
+```http
+GET /api/v1/server/help
+```
+
+**Response:** Plain text help output from `llama-server --help`
+
+### Get Llama Server Version
+
+Get version information of the llama-server binary.
+
+```http
+GET /api/v1/server/version
+```
+
+**Response:** Plain text version output from `llama-server --version`
+
+### List Available Devices
+
+List available devices for llama-server.
+
+```http
+GET /api/v1/server/devices
+```
+
+**Response:** Plain text device list from `llama-server --list-devices`
+
+## Instances
+
+### List All Instances
+
+Get a list of all instances.
+
+```http
+GET /api/v1/instances
+```
+
+**Response:**
+```json
+[
+  {
+    "name": "llama2-7b",
+    "status": "running",
+    "created": 1705312200
+  }
+]
+```
+
+### Get Instance Details
+
+Get detailed information about a specific instance.
+
+```http
+GET /api/v1/instances/{name}
+```
+
+**Response:**
+```json
+{
+  "name": "llama2-7b",
+  "status": "running",
+  "created": 1705312200
+}
+```
+
+### Create Instance
+
+Create and start a new instance.
+
+```http
+POST /api/v1/instances/{name}
+```
+
+**Request Body:** JSON object with instance configuration. Common fields include:
+
+- `backend_type`: Backend type (`llama_cpp`, `mlx_lm`, or `vllm`)
+- `backend_options`: Backend-specific configuration
+- `auto_restart`: Enable automatic restart on failure
+- `max_restarts`: Maximum restart attempts
+- `restart_delay`: Delay between restarts in seconds
+- `on_demand_start`: Start instance when receiving requests
+- `idle_timeout`: Idle timeout in minutes
+- `environment`: Environment variables as key-value pairs
+
+See [Managing Instances](managing-instances.md) for complete configuration options.
+
+**Response:**
+```json
+{
+  "name": "llama2-7b",
+  "status": "running",
+  "created": 1705312200
+}
+```
+
+### Update Instance
+
+Update an existing instance configuration. See [Managing Instances](managing-instances.md) for available configuration options.
+
+```http
+PUT /api/v1/instances/{name}
+```
+
+**Request Body:** JSON object with configuration fields to update.
+
+**Response:**
+```json
+{
+  "name": "llama2-7b",
+  "status": "running",
+  "created": 1705312200
+}
+```
+
+### Delete Instance
+
+Stop and remove an instance.
+
+```http
+DELETE /api/v1/instances/{name}
+```
+
+**Response:** `204 No Content`
+
+## Instance Operations
+
+### Start Instance
+
+Start a stopped instance.
+
+```http
+POST /api/v1/instances/{name}/start
+```
+
+**Response:**
+```json
+{
+  "name": "llama2-7b",
+  "status": "running",
+  "created": 1705312200
+}
+```
+
+**Error Responses:**
+- `409 Conflict`: Maximum number of running instances reached
+- `500 Internal Server Error`: Failed to start instance
+
+### Stop Instance
+
+Stop a running instance.
+
+```http
+POST /api/v1/instances/{name}/stop
+```
+
+**Response:**
+```json
+{
+  "name": "llama2-7b",
+  "status": "stopped",
+  "created": 1705312200
+}
+```
+
+### Restart Instance
+
+Restart an instance (stop then start).
+
+```http
+POST /api/v1/instances/{name}/restart
+```
+
+**Response:**
+```json
+{
+  "name": "llama2-7b",
+  "status": "running",
+  "created": 1705312200
+}
+```
+
+### Get Instance Logs
+
+Retrieve instance logs.
+
+```http
+GET /api/v1/instances/{name}/logs
+```
+
+**Query Parameters:**
+- `lines`: Number of lines to return (default: all lines, use -1 for all)
+
+**Response:** Plain text log output
+
+**Example:**
+```bash
+curl "http://localhost:8080/api/v1/instances/my-instance/logs?lines=100"
+```
+
+### Proxy to Instance
+
+Proxy HTTP requests directly to the llama-server instance.
+
+```http
+GET /api/v1/instances/{name}/proxy/*
+POST /api/v1/instances/{name}/proxy/*
+```
+
+This endpoint forwards all requests to the underlying llama-server instance running on its configured port. The proxy strips the `/api/v1/instances/{name}/proxy` prefix and forwards the remaining path to the instance.
+
+**Example - Check Instance Health:**
+```bash
+curl -H "Authorization: Bearer your-api-key" \
+  http://localhost:8080/api/v1/instances/my-model/proxy/health
+```
+
+This forwards the request to `http://instance-host:instance-port/health` on the actual llama-server instance.
+
+**Error Responses:**
+- `503 Service Unavailable`: Instance is not running
+
+## OpenAI-Compatible API
+
+Llamactl provides OpenAI-compatible endpoints for inference operations.
+
+### List Models
+
+List all instances in OpenAI-compatible format.
+
+```http
+GET /v1/models
+```
+
+**Response:**
+```json
+{
+  "object": "list",
+  "data": [
+    {
+      "id": "llama2-7b",
+      "object": "model",
+      "created": 1705312200,
+      "owned_by": "llamactl"
+    }
+  ]
+}
+```
+
+### Chat Completions, Completions, Embeddings
+
+All OpenAI-compatible inference endpoints are available:
+
+```http
+POST /v1/chat/completions
+POST /v1/completions
+POST /v1/embeddings
+POST /v1/rerank
+POST /v1/reranking
+```
+
+**Request Body:** Standard OpenAI format with `model` field specifying the instance name
+
+**Example:**
+```json
+{
+  "model": "llama2-7b",
+  "messages": [
+    {
+      "role": "user",
+      "content": "Hello, how are you?"
+    }
+  ]
+}
+```
+
+The server routes requests to the appropriate instance based on the `model` field in the request body. Instances with on-demand starting enabled will be automatically started if not running. For configuration details, see [Managing Instances](managing-instances.md).
+
+**Error Responses:**
+- `400 Bad Request`: Invalid request body or missing instance name
+- `503 Service Unavailable`: Instance is not running and on-demand start is disabled
+- `409 Conflict`: Cannot start instance due to maximum instances limit
+
+## Instance Status Values
+
+Instances can have the following status values:
+- `stopped`: Instance is not running
+- `running`: Instance is running and ready to accept requests
+- `failed`: Instance failed to start or crashed  
+
+## Error Responses
+
+All endpoints may return error responses in the following format:
+
+```json
+{
+  "error": "Error message description"
+}
+```
+
+### Common HTTP Status Codes
+
+- `200`: Success
+- `201`: Created
+- `204`: No Content (successful deletion)
+- `400`: Bad Request (invalid parameters or request body)
+- `401`: Unauthorized (missing or invalid API key)
+- `403`: Forbidden (insufficient permissions)
+- `404`: Not Found (instance not found)
+- `409`: Conflict (instance already exists, max instances reached)
+- `500`: Internal Server Error
+- `503`: Service Unavailable (instance not running)
+
+## Examples
+
+### Complete Instance Lifecycle
+
+```bash
+# Create and start instance
+curl -X POST http://localhost:8080/api/v1/instances/my-model \
+  -H "Content-Type: application/json" \
+  -H "Authorization: Bearer your-api-key" \
+  -d '{
+    "backend_type": "llama_cpp",
+    "backend_options": {
+      "model": "/models/llama-2-7b.gguf",
+      "gpu_layers": 32
+    },
+    "environment": {
+      "CUDA_VISIBLE_DEVICES": "0",
+      "OMP_NUM_THREADS": "8"
+    }
+  }'
+
+# Check instance status
+curl -H "Authorization: Bearer your-api-key" \
+  http://localhost:8080/api/v1/instances/my-model
+
+# Get instance logs
+curl -H "Authorization: Bearer your-api-key" \
+  "http://localhost:8080/api/v1/instances/my-model/logs?lines=50"
+
+# Use OpenAI-compatible chat completions
+curl -X POST http://localhost:8080/v1/chat/completions \
+  -H "Content-Type: application/json" \
+  -H "Authorization: Bearer your-inference-api-key" \
+  -d '{
+    "model": "my-model",
+    "messages": [
+      {"role": "user", "content": "Hello!"}
+    ],
+    "max_tokens": 100
+  }'
+
+# Stop instance
+curl -X POST -H "Authorization: Bearer your-api-key" \
+  http://localhost:8080/api/v1/instances/my-model/stop
+
+# Delete instance
+curl -X DELETE -H "Authorization: Bearer your-api-key" \
+  http://localhost:8080/api/v1/instances/my-model
+```
+
+### Using the Proxy Endpoint
+
+You can also directly proxy requests to the llama-server instance:
+
+```bash
+# Direct proxy to instance (bypasses OpenAI compatibility layer)
+curl -X POST http://localhost:8080/api/v1/instances/my-model/proxy/completion \
+  -H "Content-Type: application/json" \
+  -H "Authorization: Bearer your-api-key" \
+  -d '{
+    "prompt": "Hello, world!",
+    "n_predict": 50
+  }'
+```
+
+## Backend-Specific Endpoints
+
+### Parse Commands
+
+Llamactl provides endpoints to parse command strings from different backends into instance configuration options.
+
+#### Parse Llama.cpp Command
+
+Parse a llama-server command string into instance options.
+
+```http
+POST /api/v1/backends/llama-cpp/parse-command
+```
+
+**Request Body:**
+```json
+{
+  "command": "llama-server -m /path/to/model.gguf -c 2048 --port 8080"
+}
+```
+
+**Response:**
+```json
+{
+  "backend_type": "llama_cpp",
+  "llama_server_options": {
+    "model": "/path/to/model.gguf",
+    "ctx_size": 2048,
+    "port": 8080
+  }
+}
+```
+
+#### Parse MLX-LM Command
+
+Parse an MLX-LM server command string into instance options.
+
+```http
+POST /api/v1/backends/mlx/parse-command
+```
+
+**Request Body:**
+```json
+{
+  "command": "mlx_lm.server --model /path/to/model --port 8080"
+}
+```
+
+**Response:**
+```json
+{
+  "backend_type": "mlx_lm",
+  "mlx_server_options": {
+    "model": "/path/to/model",
+    "port": 8080
+  }
+}
+```
+
+#### Parse vLLM Command
+
+Parse a vLLM serve command string into instance options.
+
+```http
+POST /api/v1/backends/vllm/parse-command
+```
+
+**Request Body:**
+```json
+{
+  "command": "vllm serve /path/to/model --port 8080"
+}
+```
+
+**Response:**
+```json
+{
+  "backend_type": "vllm",
+  "vllm_server_options": {
+    "model": "/path/to/model",
+    "port": 8080
+  }
+}
+```
+
+**Error Responses for Parse Commands:**
+- `400 Bad Request`: Invalid request body, empty command, or parse error
+- `500 Internal Server Error`: Encoding error
+
+## Auto-Generated Documentation
+
+The API documentation is automatically generated from code annotations using Swagger/OpenAPI. To regenerate the documentation:
+
+1. Install the swag tool: `go install github.com/swaggo/swag/cmd/swag@latest`
+2. Generate docs: `swag init -g cmd/server/main.go -o apidocs`
+
+## Swagger Documentation
+
+If swagger documentation is enabled in the server configuration, you can access the interactive API documentation at:
+
+```
+http://localhost:8080/swagger/
+```
+
+This provides a complete interactive interface for testing all API endpoints.

docs/user-guide/managing-instances.md

@@ -0,0 +1,229 @@
+# Managing Instances
+
+Learn how to effectively manage your llama.cpp, MLX, and vLLM instances with Llamactl through both the Web UI and API.
+
+## Overview
+
+Llamactl provides two ways to manage instances:
+
+- **Web UI**: Accessible at `http://localhost:8080` with an intuitive dashboard
+- **REST API**: Programmatic access for automation and integration
+
+![Dashboard Screenshot](../images/dashboard.png)
+
+### Authentication
+
+If authentication is enabled:
+1. Navigate to the web UI
+2. Enter your credentials
+3. Bearer token is stored for the session
+
+### Theme Support
+
+- Switch between light and dark themes
+- Setting is remembered across sessions
+
+## Instance Cards
+
+Each instance is displayed as a card showing:
+
+- **Instance name**
+- **Health status badge** (unknown, ready, error, failed)
+- **Action buttons** (start, stop, edit, logs, delete)
+
+## Create Instance
+
+### Via Web UI
+
+![Create Instance Screenshot](../images/create_instance.png)
+
+1. Click the **"Create Instance"** button on the dashboard
+2. Enter a unique **Name** for your instance (only required field)
+3. **Choose Backend Type**:
+    - **llama.cpp**: For GGUF models using llama-server
+    - **MLX**: For MLX-optimized models (macOS only)
+    - **vLLM**: For distributed serving and high-throughput inference
+4. Configure model source:
+    - **For llama.cpp**: GGUF model path or HuggingFace repo
+    - **For MLX**: MLX model path or identifier (e.g., `mlx-community/Mistral-7B-Instruct-v0.3-4bit`)
+    - **For vLLM**: HuggingFace model identifier (e.g., `microsoft/DialoGPT-medium`)
+5. Configure optional instance management settings:
+    - **Auto Restart**: Automatically restart instance on failure
+    - **Max Restarts**: Maximum number of restart attempts
+    - **Restart Delay**: Delay in seconds between restart attempts
+    - **On Demand Start**: Start instance when receiving a request to the OpenAI compatible endpoint
+    - **Idle Timeout**: Minutes before stopping idle instance (set to 0 to disable)
+    - **Environment Variables**: Set custom environment variables for the instance process
+6. Configure backend-specific options:
+    - **llama.cpp**: Threads, context size, GPU layers, port, etc.
+    - **MLX**: Temperature, top-p, adapter path, Python environment, etc.
+    - **vLLM**: Tensor parallel size, GPU memory utilization, quantization, etc.
+7. Click **"Create"** to save the instance  
+
+### Via API
+
+```bash
+# Create llama.cpp instance with local model file
+curl -X POST http://localhost:8080/api/instances/my-llama-instance \
+  -H "Content-Type: application/json" \
+  -d '{
+    "backend_type": "llama_cpp",
+    "backend_options": {
+      "model": "/path/to/model.gguf",
+      "threads": 8,
+      "ctx_size": 4096,
+      "gpu_layers": 32
+    }
+  }'
+
+# Create MLX instance (macOS only)
+curl -X POST http://localhost:8080/api/instances/my-mlx-instance \
+  -H "Content-Type: application/json" \
+  -d '{
+    "backend_type": "mlx_lm",
+    "backend_options": {
+      "model": "mlx-community/Mistral-7B-Instruct-v0.3-4bit",
+      "temp": 0.7,
+      "top_p": 0.9,
+      "max_tokens": 2048
+    },
+    "auto_restart": true,
+    "max_restarts": 3
+  }'
+
+# Create vLLM instance
+curl -X POST http://localhost:8080/api/instances/my-vllm-instance \
+  -H "Content-Type: application/json" \
+  -d '{
+    "backend_type": "vllm",
+    "backend_options": {
+      "model": "microsoft/DialoGPT-medium",
+      "tensor_parallel_size": 2,
+      "gpu_memory_utilization": 0.9
+    },
+    "auto_restart": true,
+    "on_demand_start": true,
+    "environment": {
+      "CUDA_VISIBLE_DEVICES": "0,1",
+      "NCCL_DEBUG": "INFO",
+      "PYTHONPATH": "/custom/path"
+    }
+  }'
+
+# Create llama.cpp instance with HuggingFace model
+curl -X POST http://localhost:8080/api/instances/gemma-3-27b \
+  -H "Content-Type: application/json" \
+  -d '{
+    "backend_type": "llama_cpp",
+    "backend_options": {
+      "hf_repo": "unsloth/gemma-3-27b-it-GGUF",
+      "hf_file": "gemma-3-27b-it-GGUF.gguf",
+      "gpu_layers": 32
+    }
+  }'
+```
+
+## Start Instance
+
+### Via Web UI
+1. Click the **"Start"** button on an instance card
+2. Watch the status change to "Unknown"
+3. Monitor progress in the logs
+4. Instance status changes to "Ready" when ready
+
+### Via API
+```bash
+curl -X POST http://localhost:8080/api/instances/{name}/start
+```
+
+## Stop Instance
+
+### Via Web UI
+1. Click the **"Stop"** button on an instance card
+2. Instance gracefully shuts down
+
+### Via API
+```bash
+curl -X POST http://localhost:8080/api/instances/{name}/stop
+```
+
+## Edit Instance
+
+### Via Web UI
+1. Click the **"Edit"** button on an instance card
+2. Modify settings in the configuration dialog
+3. Changes require instance restart to take effect
+4. Click **"Update & Restart"** to apply changes
+
+### Via API
+Modify instance settings:
+
+```bash
+curl -X PUT http://localhost:8080/api/instances/{name} \
+  -H "Content-Type: application/json" \
+  -d '{
+    "backend_options": {
+      "threads": 8,
+      "context_size": 4096
+    }
+  }'
+```
+
+!!! note
+    Configuration changes require restarting the instance to take effect.
+
+
+## View Logs
+
+### Via Web UI
+
+1. Click the **"Logs"** button on any instance card
+2. Real-time log viewer opens
+
+### Via API
+Check instance status in real-time:
+
+```bash
+# Get instance details
+curl http://localhost:8080/api/instances/{name}/logs
+```
+
+## Delete Instance
+
+### Via Web UI
+1. Click the **"Delete"** button on an instance card
+2. Only stopped instances can be deleted
+3. Confirm deletion in the dialog
+
+### Via API
+```bash
+curl -X DELETE http://localhost:8080/api/instances/{name}
+```
+
+## Instance Proxy
+
+Llamactl proxies all requests to the underlying backend instances (llama-server, MLX, or vLLM).
+
+```bash
+# Get instance details
+curl http://localhost:8080/api/instances/{name}/proxy/
+```
+
+All backends provide OpenAI-compatible endpoints. Check the respective documentation:
+- [llama-server docs](https://github.com/ggml-org/llama.cpp/blob/master/tools/server/README.md)
+- [MLX-LM docs](https://github.com/ml-explore/mlx-lm/blob/main/mlx_lm/SERVER.md)
+- [vLLM docs](https://docs.vllm.ai/en/latest/)
+
+### Instance Health
+
+#### Via Web UI
+
+1. The health status badge is displayed on each instance card
+
+#### Via API
+
+Check the health status of your instances:
+
+```bash
+curl http://localhost:8080/api/instances/{name}/proxy/health
+```

docs/user-guide/troubleshooting.md

@@ -0,0 +1,160 @@
+# Troubleshooting
+
+Issues specific to Llamactl deployment and operation.
+
+## Configuration Issues
+
+### Invalid Configuration
+
+**Problem:** Invalid configuration preventing startup
+
+**Solutions:**
+1. Use minimal configuration:
+   ```yaml
+   server:
+     host: "0.0.0.0"
+     port: 8080
+   instances:
+     port_range: [8000, 9000]
+   ```
+
+2. Check data directory permissions:
+   ```bash
+   # Ensure data directory is writable (default: ~/.local/share/llamactl)
+   mkdir -p ~/.local/share/llamactl/{instances,logs}
+   ```
+
+## Instance Management Issues
+
+### Model Loading Failures
+
+**Problem:** Instance fails to start with model loading errors
+
+**Common Solutions:**  
+- **llama-server not found:** Ensure `llama-server` binary is in PATH  
+- **Wrong model format:** Ensure model is in GGUF format  
+- **Insufficient memory:** Use smaller model or reduce context size  
+- **Path issues:** Use absolute paths to model files  
+
+### Memory Issues
+
+**Problem:** Out of memory errors or system becomes unresponsive
+
+**Solutions:**
+1. **Reduce context size:**
+   ```json
+   {
+     "n_ctx": 1024
+   }
+   ```
+
+2. **Use quantized models:**  
+   - Try Q4_K_M instead of higher precision models  
+   - Use smaller model variants (7B instead of 13B)  
+
+### GPU Configuration
+
+**Problem:** GPU not being used effectively
+
+**Solutions:**
+1. **Configure GPU layers:**
+   ```json
+   {
+     "n_gpu_layers": 35
+   }
+   ```
+
+### Advanced Instance Issues
+
+**Problem:** Complex model loading, performance, or compatibility issues
+
+Since llamactl uses `llama-server` under the hood, many instance-related issues are actually llama.cpp issues. For advanced troubleshooting:
+
+**Resources:**  
+- **llama.cpp Documentation:** [https://github.com/ggml/llama.cpp](https://github.com/ggml/llama.cpp)  
+- **llama.cpp Issues:** [https://github.com/ggml/llama.cpp/issues](https://github.com/ggml/llama.cpp/issues)  
+- **llama.cpp Discussions:** [https://github.com/ggml/llama.cpp/discussions](https://github.com/ggml/llama.cpp/discussions)  
+
+**Testing directly with llama-server:**  
+```bash
+# Test your model and parameters directly with llama-server
+llama-server --model /path/to/model.gguf --port 8081 --n-gpu-layers 35
+```
+
+This helps determine if the issue is with llamactl or with the underlying llama.cpp/llama-server.
+
+## API and Network Issues
+
+### CORS Errors
+
+**Problem:** Web UI shows CORS errors in browser console
+
+**Solutions:**
+1. **Configure allowed origins:**
+   ```yaml
+   server:
+     allowed_origins:
+       - "http://localhost:3000"
+       - "https://yourdomain.com"
+   ```
+
+## Authentication Issues
+
+**Problem:** API requests failing with authentication errors
+
+**Solutions:**
+1. **Disable authentication temporarily:**
+   ```yaml
+   auth:
+     require_management_auth: false
+     require_inference_auth: false
+   ```
+
+2. **Configure API keys:**
+   ```yaml
+   auth:
+     management_keys:
+       - "your-management-key"
+     inference_keys:
+       - "your-inference-key"
+   ```
+
+3. **Use correct Authorization header:**
+   ```bash
+   curl -H "Authorization: Bearer your-api-key" \
+     http://localhost:8080/api/v1/instances
+   ```
+
+## Debugging and Logs
+
+### Viewing Instance Logs
+
+```bash
+# Get instance logs via API
+curl http://localhost:8080/api/v1/instances/{name}/logs
+
+# Or check log files directly
+tail -f ~/.local/share/llamactl/logs/{instance-name}.log
+```
+
+### Enable Debug Logging
+
+```bash
+export LLAMACTL_LOG_LEVEL=debug
+llamactl
+```
+
+## Getting Help
+
+When reporting issues, include:
+
+1. **System information:**
+   ```bash
+   llamactl --version
+   ```
+
+2. **Configuration file** (remove sensitive keys)
+
+3. **Relevant log output**
+
+4. **Steps to reproduce the issue**

mkdocs.yml

@@ -0,0 +1,79 @@
+site_name: Llamatl Documentation
+site_description: User documentation for Llamatl - A management tool for Llama.cpp instances
+site_author: Llamatl Team
+site_url: https://llamactl.org
+
+repo_name: lordmathis/llamactl
+repo_url: https://github.com/lordmathis/llamactl
+
+theme:
+  name: material
+  palette:
+    # Palette toggle for light mode
+    - scheme: default
+      primary: indigo
+      accent: indigo
+      toggle:
+        icon: material/brightness-7
+        name: Switch to dark mode
+    # Palette toggle for dark mode
+    - scheme: slate
+      primary: indigo
+      accent: indigo
+      toggle:
+        icon: material/brightness-4
+        name: Switch to light mode
+  features:
+    - navigation.tabs
+    - navigation.sections
+    - navigation.expand
+    - navigation.top
+    - search.highlight
+    - search.share
+    - content.code.copy
+
+markdown_extensions:
+  - pymdownx.highlight:
+      anchor_linenums: true
+  - pymdownx.inlinehilite
+  - pymdownx.snippets
+  - pymdownx.superfences
+  - admonition
+  - pymdownx.details
+  - pymdownx.tabbed:
+      alternate_style: true
+  - attr_list
+  - md_in_html
+  - toc:
+      permalink: true
+
+nav:
+  - Home: index.md
+  - Getting Started:
+    - Installation: getting-started/installation.md
+    - Quick Start: getting-started/quick-start.md
+    - Configuration: getting-started/configuration.md
+  - User Guide:
+    - Managing Instances: user-guide/managing-instances.md
+    - API Reference: user-guide/api-reference.md
+    - Troubleshooting: user-guide/troubleshooting.md
+
+plugins:
+  - search
+  - git-revision-date-localized
+  - mike:
+      version_selector: true
+      css_dir: css
+      javascript_dir: js
+      canonical_version: null
+
+hooks:
+  - docs/readme_sync.py
+
+extra:
+  version:
+    provider: mike
+    default: stable
+  social:
+    - icon: fontawesome/brands/github
+      link: https://github.com/lordmathis/llamactl

pkg/backends/backend.go

@@ -0,0 +1,10 @@
+package backends
+
+type BackendType string
+
+const (
+	BackendTypeLlamaCpp BackendType = "llama_cpp"
+	BackendTypeMlxLm    BackendType = "mlx_lm"
+	BackendTypeVllm     BackendType = "vllm"
+	// BackendTypeMlxVlm BackendType = "mlx_vlm"  // Future expansion
+)

pkg/backends/builder.go

@@ -0,0 +1,93 @@
+package backends
+
+import (
+	"fmt"
+	"llamactl/pkg/config"
+	"reflect"
+	"strconv"
+	"strings"
+)
+
+// BuildCommandArgs converts a struct to command line arguments
+func BuildCommandArgs(options any, multipleFlags map[string]bool) []string {
+	var args []string
+
+	v := reflect.ValueOf(options).Elem()
+	t := v.Type()
+
+	for i := 0; i < v.NumField(); i++ {
+		field := v.Field(i)
+		fieldType := t.Field(i)
+
+		if !field.CanInterface() {
+			continue
+		}
+
+		jsonTag := fieldType.Tag.Get("json")
+		if jsonTag == "" || jsonTag == "-" {
+			continue
+		}
+
+		// Get flag name from JSON tag
+		flagName := strings.Split(jsonTag, ",")[0]
+		flagName = strings.ReplaceAll(flagName, "_", "-")
+
+		switch field.Kind() {
+		case reflect.Bool:
+			if field.Bool() {
+				args = append(args, "--"+flagName)
+			}
+		case reflect.Int:
+			if field.Int() != 0 {
+				args = append(args, "--"+flagName, strconv.FormatInt(field.Int(), 10))
+			}
+		case reflect.Float64:
+			if field.Float() != 0 {
+				args = append(args, "--"+flagName, strconv.FormatFloat(field.Float(), 'f', -1, 64))
+			}
+		case reflect.String:
+			if field.String() != "" {
+				args = append(args, "--"+flagName, field.String())
+			}
+		case reflect.Slice:
+			if field.Type().Elem().Kind() == reflect.String && field.Len() > 0 {
+				if multipleFlags[flagName] {
+					// Multiple flags: --flag value1 --flag value2
+					for j := 0; j < field.Len(); j++ {
+						args = append(args, "--"+flagName, field.Index(j).String())
+					}
+				} else {
+					// Comma-separated: --flag value1,value2
+					var values []string
+					for j := 0; j < field.Len(); j++ {
+						values = append(values, field.Index(j).String())
+					}
+					args = append(args, "--"+flagName, strings.Join(values, ","))
+				}
+			}
+		}
+	}
+
+	return args
+}
+
+// BuildDockerCommand builds a Docker command with the specified configuration and arguments
+func BuildDockerCommand(backendConfig *config.BackendSettings, instanceArgs []string) (string, []string, error) {
+	// Start with configured Docker arguments (should include "run", "--rm", etc.)
+	dockerArgs := make([]string, len(backendConfig.Docker.Args))
+	copy(dockerArgs, backendConfig.Docker.Args)
+
+	// Add environment variables
+	for key, value := range backendConfig.Docker.Environment {
+		dockerArgs = append(dockerArgs, "-e", fmt.Sprintf("%s=%s", key, value))
+	}
+
+	// Add image name
+	dockerArgs = append(dockerArgs, backendConfig.Docker.Image)
+
+	// Add backend args and instance args
+	dockerArgs = append(dockerArgs, backendConfig.Args...)
+	dockerArgs = append(dockerArgs, instanceArgs...)
+
+	return "docker", dockerArgs, nil
+}

pkg/backends/llamacpp/llama.go

@@ -2,11 +2,33 @@ package llamacpp
 
 import (
 	"encoding/json"
+	"llamactl/pkg/backends"
 	"reflect"
 	"strconv"
-	"strings"
 )
 
+// multiValuedFlags defines flags that should be repeated for each value rather than comma-separated
+// Used for both parsing (with underscores) and building (with dashes)
+var multiValuedFlags = map[string]bool{
+	// Parsing keys (with underscores)
+	"override_tensor":       true,
+	"override_kv":           true,
+	"lora":                  true,
+	"lora_scaled":           true,
+	"control_vector":        true,
+	"control_vector_scaled": true,
+	"dry_sequence_breaker":  true,
+	"logit_bias":            true,
+	// Building keys (with dashes)
+	"override-tensor":       true,
+	"override-kv":           true,
+	"lora-scaled":           true,
+	"control-vector":        true,
+	"control-vector-scaled": true,
+	"dry-sequence-breaker":  true,
+	"logit-bias":            true,
+}
+
 type LlamaServerOptions struct {
 	// Common params
 	VerbosePrompt           bool     `json:"verbose_prompt,omitempty"`
@@ -315,62 +337,31 @@ func (o *LlamaServerOptions) UnmarshalJSON(data []byte) error {
 
 // BuildCommandArgs converts InstanceOptions to command line arguments
 func (o *LlamaServerOptions) BuildCommandArgs() []string {
-	var args []string
+	// Llama uses multiple flags for arrays by default (not comma-separated)
+	// Use package-level multiValuedFlags variable
+	return backends.BuildCommandArgs(o, multiValuedFlags)
+}
 
-	v := reflect.ValueOf(o).Elem()
-	t := v.Type()
+func (o *LlamaServerOptions) BuildDockerArgs() []string {
+	// For llama, Docker args are the same as normal args
+	return o.BuildCommandArgs()
+}
 
-	for i := 0; i < v.NumField(); i++ {
-		field := v.Field(i)
-		fieldType := t.Field(i)
+// ParseLlamaCommand parses a llama-server command string into LlamaServerOptions
+// Supports multiple formats:
+// 1. Full command: "llama-server --model file.gguf"
+// 2. Full path: "/usr/local/bin/llama-server --model file.gguf"
+// 3. Args only: "--model file.gguf --gpu-layers 32"
+// 4. Multiline commands with backslashes
+func ParseLlamaCommand(command string) (*LlamaServerOptions, error) {
+	executableNames := []string{"llama-server"}
+	var subcommandNames []string // Llama has no subcommands
+	// Use package-level multiValuedFlags variable
 
-		// Skip unexported fields
-		if !field.CanInterface() {
-			continue
-		}
-
-		// Get the JSON tag to determine the flag name
-		jsonTag := fieldType.Tag.Get("json")
-		if jsonTag == "" || jsonTag == "-" {
-			continue
-		}
-
-		// Remove ",omitempty" from the tag
-		flagName := jsonTag
-		if commaIndex := strings.Index(jsonTag, ","); commaIndex != -1 {
-			flagName = jsonTag[:commaIndex]
-		}
-
-		// Convert snake_case to kebab-case for CLI flags
-		flagName = strings.ReplaceAll(flagName, "_", "-")
-
-		// Add the appropriate arguments based on field type and value
-		switch field.Kind() {
-		case reflect.Bool:
-			if field.Bool() {
-				args = append(args, "--"+flagName)
-			}
-		case reflect.Int:
-			if field.Int() != 0 {
-				args = append(args, "--"+flagName, strconv.FormatInt(field.Int(), 10))
-			}
-		case reflect.Float64:
-			if field.Float() != 0 {
-				args = append(args, "--"+flagName, strconv.FormatFloat(field.Float(), 'f', -1, 64))
-			}
-		case reflect.String:
-			if field.String() != "" {
-				args = append(args, "--"+flagName, field.String())
-			}
-		case reflect.Slice:
-			if field.Type().Elem().Kind() == reflect.String {
-				// Handle []string fields
-				for j := 0; j < field.Len(); j++ {
-					args = append(args, "--"+flagName, field.Index(j).String())
-				}
-			}
-		}
+	var llamaOptions LlamaServerOptions
+	if err := backends.ParseCommand(command, executableNames, subcommandNames, multiValuedFlags, &llamaOptions); err != nil {
+		return nil, err
 	}
 
-	return args
+	return &llamaOptions, nil
 }

pkg/backends/llamacpp/llama_test.go

@@ -378,6 +378,121 @@ func TestUnmarshalJSON_ArrayFields(t *testing.T) {
 	}
 }
 
+func TestParseLlamaCommand(t *testing.T) {
+	tests := []struct {
+		name      string
+		command   string
+		expectErr bool
+	}{
+		{
+			name:      "basic command",
+			command:   "llama-server --model /path/to/model.gguf --gpu-layers 32",
+			expectErr: false,
+		},
+		{
+			name:      "args only",
+			command:   "--model /path/to/model.gguf --ctx-size 4096",
+			expectErr: false,
+		},
+		{
+			name:      "mixed flag formats",
+			command:   "llama-server --model=/path/model.gguf --gpu-layers 16 --verbose",
+			expectErr: false,
+		},
+		{
+			name:      "quoted strings",
+			command:   `llama-server --model test.gguf --api-key "sk-1234567890abcdef"`,
+			expectErr: false,
+		},
+		{
+			name:      "empty command",
+			command:   "",
+			expectErr: true,
+		},
+		{
+			name:      "unterminated quote",
+			command:   `llama-server --model test.gguf --api-key "unterminated`,
+			expectErr: true,
+		},
+		{
+			name:      "malformed flag",
+			command:   "llama-server ---model test.gguf",
+			expectErr: true,
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			result, err := llamacpp.ParseLlamaCommand(tt.command)
+
+			if tt.expectErr {
+				if err == nil {
+					t.Errorf("expected error but got none")
+				}
+				return
+			}
+
+			if err != nil {
+				t.Errorf("unexpected error: %v", err)
+				return
+			}
+
+			if result == nil {
+				t.Errorf("expected result but got nil")
+			}
+		})
+	}
+}
+
+func TestParseLlamaCommandValues(t *testing.T) {
+	command := "llama-server --model /test/model.gguf --gpu-layers 32 --temp 0.7 --verbose --no-mmap"
+	result, err := llamacpp.ParseLlamaCommand(command)
+
+	if err != nil {
+		t.Fatalf("unexpected error: %v", err)
+	}
+
+	if result.Model != "/test/model.gguf" {
+		t.Errorf("expected model '/test/model.gguf', got '%s'", result.Model)
+	}
+
+	if result.GPULayers != 32 {
+		t.Errorf("expected gpu_layers 32, got %d", result.GPULayers)
+	}
+
+	if result.Temperature != 0.7 {
+		t.Errorf("expected temperature 0.7, got %f", result.Temperature)
+	}
+
+	if !result.Verbose {
+		t.Errorf("expected verbose to be true")
+	}
+
+	if !result.NoMmap {
+		t.Errorf("expected no_mmap to be true")
+	}
+}
+
+func TestParseLlamaCommandArrays(t *testing.T) {
+	command := "llama-server --model test.gguf --lora adapter1.bin --lora=adapter2.bin"
+	result, err := llamacpp.ParseLlamaCommand(command)
+
+	if err != nil {
+		t.Fatalf("unexpected error: %v", err)
+	}
+
+	if len(result.Lora) != 2 {
+		t.Errorf("expected 2 lora adapters, got %d", len(result.Lora))
+	}
+
+	expected := []string{"adapter1.bin", "adapter2.bin"}
+	for i, v := range expected {
+		if result.Lora[i] != v {
+			t.Errorf("expected lora[%d]=%s got %s", i, v, result.Lora[i])
+		}
+	}
+}
+
 // Helper functions
 func contains(slice []string, item string) bool {
 	return slices.Contains(slice, item)

pkg/backends/mlx/mlx.go

@@ -0,0 +1,56 @@
+package mlx
+
+import (
+	"llamactl/pkg/backends"
+)
+
+type MlxServerOptions struct {
+	// Basic connection options
+	Model string `json:"model,omitempty"`
+	Host  string `json:"host,omitempty"`
+	Port  int    `json:"port,omitempty"`
+
+	// Model and adapter options
+	AdapterPath     string `json:"adapter_path,omitempty"`
+	DraftModel      string `json:"draft_model,omitempty"`
+	NumDraftTokens  int    `json:"num_draft_tokens,omitempty"`
+	TrustRemoteCode bool   `json:"trust_remote_code,omitempty"`
+
+	// Logging and templates
+	LogLevel               string `json:"log_level,omitempty"`
+	ChatTemplate           string `json:"chat_template,omitempty"`
+	UseDefaultChatTemplate bool   `json:"use_default_chat_template,omitempty"`
+	ChatTemplateArgs       string `json:"chat_template_args,omitempty"` // JSON string
+
+	// Sampling defaults
+	Temp      float64 `json:"temp,omitempty"`
+	TopP      float64 `json:"top_p,omitempty"`
+	TopK      int     `json:"top_k,omitempty"`
+	MinP      float64 `json:"min_p,omitempty"`
+	MaxTokens int     `json:"max_tokens,omitempty"`
+}
+
+// BuildCommandArgs converts to command line arguments
+func (o *MlxServerOptions) BuildCommandArgs() []string {
+	multipleFlags := map[string]bool{} // MLX doesn't currently have []string fields
+	return backends.BuildCommandArgs(o, multipleFlags)
+}
+
+// ParseMlxCommand parses a mlx_lm.server command string into MlxServerOptions
+// Supports multiple formats:
+// 1. Full command: "mlx_lm.server --model model/path"
+// 2. Full path: "/usr/local/bin/mlx_lm.server --model model/path"
+// 3. Args only: "--model model/path --host 0.0.0.0"
+// 4. Multiline commands with backslashes
+func ParseMlxCommand(command string) (*MlxServerOptions, error) {
+	executableNames := []string{"mlx_lm.server"}
+	var subcommandNames []string          // MLX has no subcommands
+	multiValuedFlags := map[string]bool{} // MLX has no multi-valued flags
+
+	var mlxOptions MlxServerOptions
+	if err := backends.ParseCommand(command, executableNames, subcommandNames, multiValuedFlags, &mlxOptions); err != nil {
+		return nil, err
+	}
+
+	return &mlxOptions, nil
+}

pkg/backends/mlx/mlx_test.go

@@ -0,0 +1,157 @@
+package mlx_test
+
+import (
+	"llamactl/pkg/backends/mlx"
+	"testing"
+)
+
+func TestParseMlxCommand(t *testing.T) {
+	tests := []struct {
+		name      string
+		command   string
+		expectErr bool
+	}{
+		{
+			name:      "basic command",
+			command:   "mlx_lm.server --model /path/to/model --host 0.0.0.0",
+			expectErr: false,
+		},
+		{
+			name:      "args only",
+			command:   "--model /path/to/model --port 8080",
+			expectErr: false,
+		},
+		{
+			name:      "mixed flag formats",
+			command:   "mlx_lm.server --model=/path/model --temp=0.7 --trust-remote-code",
+			expectErr: false,
+		},
+		{
+			name:      "quoted strings",
+			command:   `mlx_lm.server --model test.mlx --chat-template "User: {user}\nAssistant: "`,
+			expectErr: false,
+		},
+		{
+			name:      "empty command",
+			command:   "",
+			expectErr: true,
+		},
+		{
+			name:      "unterminated quote",
+			command:   `mlx_lm.server --model test.mlx --chat-template "unterminated`,
+			expectErr: true,
+		},
+		{
+			name:      "malformed flag",
+			command:   "mlx_lm.server ---model test.mlx",
+			expectErr: true,
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			result, err := mlx.ParseMlxCommand(tt.command)
+
+			if tt.expectErr {
+				if err == nil {
+					t.Errorf("expected error but got none")
+				}
+				return
+			}
+
+			if err != nil {
+				t.Errorf("unexpected error: %v", err)
+				return
+			}
+
+			if result == nil {
+				t.Errorf("expected result but got nil")
+			}
+		})
+	}
+}
+
+func TestParseMlxCommandValues(t *testing.T) {
+	command := "mlx_lm.server --model /test/model.mlx --port 8080 --temp 0.7 --trust-remote-code --log-level DEBUG"
+	result, err := mlx.ParseMlxCommand(command)
+
+	if err != nil {
+		t.Fatalf("unexpected error: %v", err)
+	}
+
+	if result.Model != "/test/model.mlx" {
+		t.Errorf("expected model '/test/model.mlx', got '%s'", result.Model)
+	}
+
+	if result.Port != 8080 {
+		t.Errorf("expected port 8080, got %d", result.Port)
+	}
+
+	if result.Temp != 0.7 {
+		t.Errorf("expected temp 0.7, got %f", result.Temp)
+	}
+
+	if !result.TrustRemoteCode {
+		t.Errorf("expected trust_remote_code to be true")
+	}
+
+	if result.LogLevel != "DEBUG" {
+		t.Errorf("expected log_level 'DEBUG', got '%s'", result.LogLevel)
+	}
+}
+
+func TestBuildCommandArgs(t *testing.T) {
+	options := &mlx.MlxServerOptions{
+		Model:           "/test/model.mlx",
+		Host:            "127.0.0.1",
+		Port:            8080,
+		Temp:            0.7,
+		TopP:            0.9,
+		TopK:            40,
+		MaxTokens:       2048,
+		TrustRemoteCode: true,
+		LogLevel:        "DEBUG",
+		ChatTemplate:    "custom template",
+	}
+
+	args := options.BuildCommandArgs()
+
+	// Check that all expected flags are present
+	expectedFlags := map[string]string{
+		"--model":         "/test/model.mlx",
+		"--host":          "127.0.0.1",
+		"--port":          "8080",
+		"--log-level":     "DEBUG",
+		"--chat-template": "custom template",
+		"--temp":          "0.7",
+		"--top-p":         "0.9",
+		"--top-k":         "40",
+		"--max-tokens":    "2048",
+	}
+
+	for i := 0; i < len(args); i++ {
+		if args[i] == "--trust-remote-code" {
+			continue // Boolean flag with no value
+		}
+		if args[i] == "--use-default-chat-template" {
+			continue // Boolean flag with no value
+		}
+
+		if expectedValue, exists := expectedFlags[args[i]]; exists && i+1 < len(args) {
+			if args[i+1] != expectedValue {
+				t.Errorf("expected %s to have value %s, got %s", args[i], expectedValue, args[i+1])
+			}
+		}
+	}
+
+	// Check boolean flags
+	foundTrustRemoteCode := false
+	for _, arg := range args {
+		if arg == "--trust-remote-code" {
+			foundTrustRemoteCode = true
+		}
+	}
+	if !foundTrustRemoteCode {
+		t.Errorf("expected --trust-remote-code flag to be present")
+	}
+}

pkg/backends/parser.go

@@ -0,0 +1,213 @@
+package backends
+
+import (
+	"encoding/json"
+	"fmt"
+	"path/filepath"
+	"regexp"
+	"strconv"
+	"strings"
+)
+
+// ParseCommand parses a command string into a target struct
+func ParseCommand(command string, executableNames []string, subcommandNames []string, multiValuedFlags map[string]bool, target any) error {
+	// Normalize multiline commands
+	command = normalizeCommand(command)
+	if command == "" {
+		return fmt.Errorf("command cannot be empty")
+	}
+
+	// Extract arguments and positional model
+	args, modelFromPositional, err := extractArgs(command, executableNames, subcommandNames)
+	if err != nil {
+		return err
+	}
+
+	// Parse flags into map
+	options, err := parseFlags(args, multiValuedFlags)
+	if err != nil {
+		return err
+	}
+
+	// If we found a positional model and no --model flag was provided, set the model
+	if modelFromPositional != "" {
+		if _, hasModelFlag := options["model"]; !hasModelFlag {
+			options["model"] = modelFromPositional
+		}
+	}
+
+	// Convert to target struct via JSON
+	jsonData, err := json.Marshal(options)
+	if err != nil {
+		return fmt.Errorf("failed to marshal options: %w", err)
+	}
+
+	if err := json.Unmarshal(jsonData, target); err != nil {
+		return fmt.Errorf("failed to unmarshal to target: %w", err)
+	}
+
+	return nil
+}
+
+// normalizeCommand handles multiline commands with backslashes
+func normalizeCommand(command string) string {
+	re := regexp.MustCompile(`\\\s*\n\s*`)
+	normalized := re.ReplaceAllString(command, " ")
+	re = regexp.MustCompile(`\s+`)
+	return strings.TrimSpace(re.ReplaceAllString(normalized, " "))
+}
+
+// extractArgs extracts arguments from command, removing executable and subcommands
+// Returns: args, modelFromPositional, error
+func extractArgs(command string, executableNames []string, subcommandNames []string) ([]string, string, error) {
+	// Check for unterminated quotes
+	if strings.Count(command, `"`)%2 != 0 || strings.Count(command, `'`)%2 != 0 {
+		return nil, "", fmt.Errorf("unterminated quoted string")
+	}
+
+	tokens := strings.Fields(command)
+	if len(tokens) == 0 {
+		return nil, "", fmt.Errorf("no tokens found")
+	}
+
+	// Skip executable
+	start := 0
+	firstToken := tokens[0]
+
+	// Check for executable name (with or without path)
+	if strings.Contains(firstToken, string(filepath.Separator)) {
+		baseName := filepath.Base(firstToken)
+		for _, execName := range executableNames {
+			if strings.HasSuffix(strings.ToLower(baseName), strings.ToLower(execName)) {
+				start = 1
+				break
+			}
+		}
+	} else {
+		for _, execName := range executableNames {
+			if strings.EqualFold(firstToken, execName) {
+				start = 1
+				break
+			}
+		}
+	}
+
+	// Skip subcommand if present
+	if start < len(tokens) {
+		for _, subCmd := range subcommandNames {
+			if strings.EqualFold(tokens[start], subCmd) {
+				start++
+				break
+			}
+		}
+	}
+
+	// Handle case where command starts with subcommand (no executable)
+	if start == 0 {
+		for _, subCmd := range subcommandNames {
+			if strings.EqualFold(firstToken, subCmd) {
+				start = 1
+				break
+			}
+		}
+	}
+
+	args := tokens[start:]
+
+	// Extract first positional argument (model) if present and not a flag
+	var modelFromPositional string
+	if len(args) > 0 && !strings.HasPrefix(args[0], "-") {
+		modelFromPositional = args[0]
+		args = args[1:] // Remove the model from args to process remaining flags
+	}
+
+	return args, modelFromPositional, nil
+}
+
+// parseFlags parses command line flags into a map
+func parseFlags(args []string, multiValuedFlags map[string]bool) (map[string]any, error) {
+	options := make(map[string]any)
+
+	for i := 0; i < len(args); i++ {
+		arg := args[i]
+
+		if !strings.HasPrefix(arg, "-") {
+			continue
+		}
+
+		// Check for malformed flags (more than two leading dashes)
+		if strings.HasPrefix(arg, "---") {
+			return nil, fmt.Errorf("malformed flag: %s", arg)
+		}
+
+		// Get flag name and value
+		var flagName, value string
+		var hasValue bool
+
+		if strings.Contains(arg, "=") {
+			parts := strings.SplitN(arg, "=", 2)
+			flagName = strings.TrimLeft(parts[0], "-")
+			value = parts[1]
+			hasValue = true
+		} else {
+			flagName = strings.TrimLeft(arg, "-")
+			if i+1 < len(args) && !strings.HasPrefix(args[i+1], "-") {
+				value = args[i+1]
+				hasValue = true
+				i++ // Skip next arg since we consumed it
+			}
+		}
+
+		// Convert kebab-case to snake_case for JSON
+		flagName = strings.ReplaceAll(flagName, "-", "_")
+
+		if hasValue {
+			// Handle multi-valued flags
+			if multiValuedFlags[flagName] {
+				if existing, ok := options[flagName].([]string); ok {
+					options[flagName] = append(existing, value)
+				} else {
+					options[flagName] = []string{value}
+				}
+			} else {
+				options[flagName] = parseValue(value)
+			}
+		} else {
+			// Boolean flag
+			options[flagName] = true
+		}
+	}
+
+	return options, nil
+}
+
+// parseValue converts string to appropriate type
+func parseValue(value string) any {
+	// Remove quotes
+	if len(value) >= 2 {
+		if (value[0] == '"' && value[len(value)-1] == '"') || (value[0] == '\'' && value[len(value)-1] == '\'') {
+			value = value[1 : len(value)-1]
+		}
+	}
+
+	// Try boolean
+	switch strings.ToLower(value) {
+	case "true":
+		return true
+	case "false":
+		return false
+	}
+
+	// Try integer
+	if intVal, err := strconv.Atoi(value); err == nil {
+		return intVal
+	}
+
+	// Try float
+	if floatVal, err := strconv.ParseFloat(value, 64); err == nil {
+		return floatVal
+	}
+
+	// Return as string
+	return value
+}

pkg/backends/vllm/vllm.go

@@ -0,0 +1,200 @@
+package vllm
+
+import (
+	"llamactl/pkg/backends"
+)
+
+// multiValuedFlags defines flags that should be repeated for each value rather than comma-separated
+var multiValuedFlags = map[string]bool{
+	"api-key":         true,
+	"allowed-origins": true,
+	"allowed-methods": true,
+	"allowed-headers": true,
+	"middleware":      true,
+}
+
+type VllmServerOptions struct {
+	// Basic connection options (auto-assigned by llamactl)
+	Host string `json:"host,omitempty"`
+	Port int    `json:"port,omitempty"`
+
+	// Model and engine configuration
+	Model                      string `json:"model,omitempty"`
+	Tokenizer                  string `json:"tokenizer,omitempty"`
+	SkipTokenizerInit          bool   `json:"skip_tokenizer_init,omitempty"`
+	Revision                   string `json:"revision,omitempty"`
+	CodeRevision               string `json:"code_revision,omitempty"`
+	TokenizerRevision          string `json:"tokenizer_revision,omitempty"`
+	TokenizerMode              string `json:"tokenizer_mode,omitempty"`
+	TrustRemoteCode            bool   `json:"trust_remote_code,omitempty"`
+	DownloadDir                string `json:"download_dir,omitempty"`
+	LoadFormat                 string `json:"load_format,omitempty"`
+	ConfigFormat               string `json:"config_format,omitempty"`
+	Dtype                      string `json:"dtype,omitempty"`
+	KVCacheDtype               string `json:"kv_cache_dtype,omitempty"`
+	QuantizationParamPath      string `json:"quantization_param_path,omitempty"`
+	Seed                       int    `json:"seed,omitempty"`
+	MaxModelLen                int    `json:"max_model_len,omitempty"`
+	GuidedDecodingBackend      string `json:"guided_decoding_backend,omitempty"`
+	DistributedExecutorBackend string `json:"distributed_executor_backend,omitempty"`
+	WorkerUseRay               bool   `json:"worker_use_ray,omitempty"`
+	RayWorkersUseNSight        bool   `json:"ray_workers_use_nsight,omitempty"`
+
+	// Performance and serving configuration
+	BlockSize                int     `json:"block_size,omitempty"`
+	EnablePrefixCaching      bool    `json:"enable_prefix_caching,omitempty"`
+	DisableSlidingWindow     bool    `json:"disable_sliding_window,omitempty"`
+	UseV2BlockManager        bool    `json:"use_v2_block_manager,omitempty"`
+	NumLookaheadSlots        int     `json:"num_lookahead_slots,omitempty"`
+	SwapSpace                int     `json:"swap_space,omitempty"`
+	CPUOffloadGB             int     `json:"cpu_offload_gb,omitempty"`
+	GPUMemoryUtilization     float64 `json:"gpu_memory_utilization,omitempty"`
+	NumGPUBlocksOverride     int     `json:"num_gpu_blocks_override,omitempty"`
+	MaxNumBatchedTokens      int     `json:"max_num_batched_tokens,omitempty"`
+	MaxNumSeqs               int     `json:"max_num_seqs,omitempty"`
+	MaxLogprobs              int     `json:"max_logprobs,omitempty"`
+	DisableLogStats          bool    `json:"disable_log_stats,omitempty"`
+	Quantization             string  `json:"quantization,omitempty"`
+	RopeScaling              string  `json:"rope_scaling,omitempty"`
+	RopeTheta                float64 `json:"rope_theta,omitempty"`
+	EnforceEager             bool    `json:"enforce_eager,omitempty"`
+	MaxContextLenToCapture   int     `json:"max_context_len_to_capture,omitempty"`
+	MaxSeqLenToCapture       int     `json:"max_seq_len_to_capture,omitempty"`
+	DisableCustomAllReduce   bool    `json:"disable_custom_all_reduce,omitempty"`
+	TokenizerPoolSize        int     `json:"tokenizer_pool_size,omitempty"`
+	TokenizerPoolType        string  `json:"tokenizer_pool_type,omitempty"`
+	TokenizerPoolExtraConfig string  `json:"tokenizer_pool_extra_config,omitempty"`
+	EnableLoraBias           bool    `json:"enable_lora_bias,omitempty"`
+	LoraExtraVocabSize       int     `json:"lora_extra_vocab_size,omitempty"`
+	LoraRank                 int     `json:"lora_rank,omitempty"`
+	PromptLookbackDistance   int     `json:"prompt_lookback_distance,omitempty"`
+	PreemptionMode           string  `json:"preemption_mode,omitempty"`
+
+	// Distributed and parallel processing
+	TensorParallelSize            int     `json:"tensor_parallel_size,omitempty"`
+	PipelineParallelSize          int     `json:"pipeline_parallel_size,omitempty"`
+	MaxParallelLoadingWorkers     int     `json:"max_parallel_loading_workers,omitempty"`
+	DisableAsyncOutputProc        bool    `json:"disable_async_output_proc,omitempty"`
+	WorkerClass                   string  `json:"worker_class,omitempty"`
+	EnabledLoraModules            string  `json:"enabled_lora_modules,omitempty"`
+	MaxLoraRank                   int     `json:"max_lora_rank,omitempty"`
+	FullyShardedLoras             bool    `json:"fully_sharded_loras,omitempty"`
+	LoraModules                   string  `json:"lora_modules,omitempty"`
+	PromptAdapters                string  `json:"prompt_adapters,omitempty"`
+	MaxPromptAdapterToken         int     `json:"max_prompt_adapter_token,omitempty"`
+	Device                        string  `json:"device,omitempty"`
+	SchedulerDelay                float64 `json:"scheduler_delay,omitempty"`
+	EnableChunkedPrefill          bool    `json:"enable_chunked_prefill,omitempty"`
+	SpeculativeModel              string  `json:"speculative_model,omitempty"`
+	SpeculativeModelQuantization  string  `json:"speculative_model_quantization,omitempty"`
+	SpeculativeRevision           string  `json:"speculative_revision,omitempty"`
+	SpeculativeMaxModelLen        int     `json:"speculative_max_model_len,omitempty"`
+	SpeculativeDisableByBatchSize int     `json:"speculative_disable_by_batch_size,omitempty"`
+	NgptSpeculativeLength         int     `json:"ngpt_speculative_length,omitempty"`
+	SpeculativeDisableMqa         bool    `json:"speculative_disable_mqa,omitempty"`
+	ModelLoaderExtraConfig        string  `json:"model_loader_extra_config,omitempty"`
+	IgnorePatterns                string  `json:"ignore_patterns,omitempty"`
+	PreloadedLoraModules          string  `json:"preloaded_lora_modules,omitempty"`
+
+	// OpenAI server specific options
+	UDS                            string   `json:"uds,omitempty"`
+	UvicornLogLevel                string   `json:"uvicorn_log_level,omitempty"`
+	ResponseRole                   string   `json:"response_role,omitempty"`
+	SSLKeyfile                     string   `json:"ssl_keyfile,omitempty"`
+	SSLCertfile                    string   `json:"ssl_certfile,omitempty"`
+	SSLCACerts                     string   `json:"ssl_ca_certs,omitempty"`
+	SSLCertReqs                    int      `json:"ssl_cert_reqs,omitempty"`
+	RootPath                       string   `json:"root_path,omitempty"`
+	Middleware                     []string `json:"middleware,omitempty"`
+	ReturnTokensAsTokenIDS         bool     `json:"return_tokens_as_token_ids,omitempty"`
+	DisableFrontendMultiprocessing bool     `json:"disable_frontend_multiprocessing,omitempty"`
+	EnableAutoToolChoice           bool     `json:"enable_auto_tool_choice,omitempty"`
+	ToolCallParser                 string   `json:"tool_call_parser,omitempty"`
+	ToolServer                     string   `json:"tool_server,omitempty"`
+	ChatTemplate                   string   `json:"chat_template,omitempty"`
+	ChatTemplateContentFormat      string   `json:"chat_template_content_format,omitempty"`
+	AllowCredentials               bool     `json:"allow_credentials,omitempty"`
+	AllowedOrigins                 []string `json:"allowed_origins,omitempty"`
+	AllowedMethods                 []string `json:"allowed_methods,omitempty"`
+	AllowedHeaders                 []string `json:"allowed_headers,omitempty"`
+	APIKey                         []string `json:"api_key,omitempty"`
+	EnableLogOutputs               bool     `json:"enable_log_outputs,omitempty"`
+	EnableTokenUsage               bool     `json:"enable_token_usage,omitempty"`
+	EnableAsyncEngineDebug         bool     `json:"enable_async_engine_debug,omitempty"`
+	EngineUseRay                   bool     `json:"engine_use_ray,omitempty"`
+	DisableLogRequests             bool     `json:"disable_log_requests,omitempty"`
+	MaxLogLen                      int      `json:"max_log_len,omitempty"`
+
+	// Additional engine configuration
+	Task                      string `json:"task,omitempty"`
+	MultiModalConfig          string `json:"multi_modal_config,omitempty"`
+	LimitMmPerPrompt          string `json:"limit_mm_per_prompt,omitempty"`
+	EnableSleepMode           bool   `json:"enable_sleep_mode,omitempty"`
+	EnableChunkingRequest     bool   `json:"enable_chunking_request,omitempty"`
+	CompilationConfig         string `json:"compilation_config,omitempty"`
+	DisableSlidingWindowMask  bool   `json:"disable_sliding_window_mask,omitempty"`
+	EnableTRTLLMEngineLatency bool   `json:"enable_trtllm_engine_latency,omitempty"`
+	OverridePoolingConfig     string `json:"override_pooling_config,omitempty"`
+	OverrideNeuronConfig      string `json:"override_neuron_config,omitempty"`
+	OverrideKVCacheALIGNSize  int    `json:"override_kv_cache_align_size,omitempty"`
+}
+
+// BuildCommandArgs converts VllmServerOptions to command line arguments
+// For vLLM native, model is a positional argument after "serve"
+func (o *VllmServerOptions) BuildCommandArgs() []string {
+	var args []string
+
+	// Add model as positional argument if specified (for native execution)
+	if o.Model != "" {
+		args = append(args, o.Model)
+	}
+
+	// Create a copy without Model field to avoid --model flag
+	optionsCopy := *o
+	optionsCopy.Model = ""
+
+	// Use package-level multipleFlags variable
+
+	flagArgs := backends.BuildCommandArgs(&optionsCopy, multiValuedFlags)
+	args = append(args, flagArgs...)
+
+	return args
+}
+
+func (o *VllmServerOptions) BuildDockerArgs() []string {
+	var args []string
+
+	// Use package-level multipleFlags variable
+	flagArgs := backends.BuildCommandArgs(o, multiValuedFlags)
+	args = append(args, flagArgs...)
+
+	return args
+}
+
+// ParseVllmCommand parses a vLLM serve command string into VllmServerOptions
+// Supports multiple formats:
+// 1. Full command: "vllm serve --model MODEL_NAME --other-args"
+// 2. Full path: "/usr/local/bin/vllm serve --model MODEL_NAME"
+// 3. Serve only: "serve --model MODEL_NAME --other-args"
+// 4. Args only: "--model MODEL_NAME --other-args"
+// 5. Multiline commands with backslashes
+func ParseVllmCommand(command string) (*VllmServerOptions, error) {
+	executableNames := []string{"vllm"}
+	subcommandNames := []string{"serve"}
+	multiValuedFlags := map[string]bool{
+		"middleware":      true,
+		"api_key":         true,
+		"allowed_origins": true,
+		"allowed_methods": true,
+		"allowed_headers": true,
+		"lora_modules":    true,
+		"prompt_adapters": true,
+	}
+
+	var vllmOptions VllmServerOptions
+	if err := backends.ParseCommand(command, executableNames, subcommandNames, multiValuedFlags, &vllmOptions); err != nil {
+		return nil, err
+	}
+
+	return &vllmOptions, nil
+}

pkg/backends/vllm/vllm_test.go

@@ -0,0 +1,153 @@
+package vllm_test
+
+import (
+	"llamactl/pkg/backends/vllm"
+	"slices"
+	"testing"
+)
+
+func TestParseVllmCommand(t *testing.T) {
+	tests := []struct {
+		name      string
+		command   string
+		expectErr bool
+	}{
+		{
+			name:      "basic vllm serve command",
+			command:   "vllm serve microsoft/DialoGPT-medium",
+			expectErr: false,
+		},
+		{
+			name:      "serve only command",
+			command:   "serve microsoft/DialoGPT-medium",
+			expectErr: false,
+		},
+		{
+			name:      "positional model with flags",
+			command:   "vllm serve microsoft/DialoGPT-medium --tensor-parallel-size 2",
+			expectErr: false,
+		},
+		{
+			name:      "model with path",
+			command:   "vllm serve /path/to/model --gpu-memory-utilization 0.8",
+			expectErr: false,
+		},
+		{
+			name:      "empty command",
+			command:   "",
+			expectErr: true,
+		},
+		{
+			name:      "unterminated quote",
+			command:   `vllm serve "unterminated`,
+			expectErr: true,
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			result, err := vllm.ParseVllmCommand(tt.command)
+
+			if tt.expectErr {
+				if err == nil {
+					t.Errorf("expected error but got none")
+				}
+				return
+			}
+
+			if err != nil {
+				t.Errorf("unexpected error: %v", err)
+				return
+			}
+
+			if result == nil {
+				t.Errorf("expected result but got nil")
+			}
+		})
+	}
+}
+
+func TestParseVllmCommandValues(t *testing.T) {
+	command := "vllm serve test-model --tensor-parallel-size 4 --gpu-memory-utilization 0.8 --enable-log-outputs"
+	result, err := vllm.ParseVllmCommand(command)
+
+	if err != nil {
+		t.Fatalf("unexpected error: %v", err)
+	}
+
+	if result.Model != "test-model" {
+		t.Errorf("expected model 'test-model', got '%s'", result.Model)
+	}
+	if result.TensorParallelSize != 4 {
+		t.Errorf("expected tensor_parallel_size 4, got %d", result.TensorParallelSize)
+	}
+	if result.GPUMemoryUtilization != 0.8 {
+		t.Errorf("expected gpu_memory_utilization 0.8, got %f", result.GPUMemoryUtilization)
+	}
+	if !result.EnableLogOutputs {
+		t.Errorf("expected enable_log_outputs true, got %v", result.EnableLogOutputs)
+	}
+}
+
+func TestBuildCommandArgs(t *testing.T) {
+	options := vllm.VllmServerOptions{
+		Model:                "microsoft/DialoGPT-medium",
+		Port:                 8080,
+		Host:                 "localhost",
+		TensorParallelSize:   2,
+		GPUMemoryUtilization: 0.8,
+		EnableLogOutputs:     true,
+		AllowedOrigins:       []string{"http://localhost:3000", "https://example.com"},
+	}
+
+	args := options.BuildCommandArgs()
+
+	// Check that model is the first positional argument (not a --model flag)
+	if len(args) == 0 || args[0] != "microsoft/DialoGPT-medium" {
+		t.Errorf("Expected model 'microsoft/DialoGPT-medium' as first positional argument, got args: %v", args)
+	}
+
+	// Check that --model flag is NOT present (since model should be positional)
+	if contains(args, "--model") {
+		t.Errorf("Found --model flag, but model should be positional argument in args: %v", args)
+	}
+
+	// Check other flags
+	if !containsFlagWithValue(args, "--tensor-parallel-size", "2") {
+		t.Errorf("Expected --tensor-parallel-size 2 not found in %v", args)
+	}
+	if !contains(args, "--enable-log-outputs") {
+		t.Errorf("Expected --enable-log-outputs not found in %v", args)
+	}
+	if !contains(args, "--host") {
+		t.Errorf("Expected --host not found in %v", args)
+	}
+	if !contains(args, "--port") {
+		t.Errorf("Expected --port not found in %v", args)
+	}
+
+	// Check array handling (multiple flags)
+	allowedOriginsCount := 0
+	for i := range args {
+		if args[i] == "--allowed-origins" {
+			allowedOriginsCount++
+		}
+	}
+	if allowedOriginsCount != 2 {
+		t.Errorf("Expected 2 --allowed-origins flags, got %d", allowedOriginsCount)
+	}
+}
+
+// Helper functions
+func contains(slice []string, item string) bool {
+	return slices.Contains(slice, item)
+}
+
+func containsFlagWithValue(args []string, flag, value string) bool {
+	for i, arg := range args {
+		if arg == flag && i+1 < len(args) && args[i+1] == value {
+			return true
+		}
+	}
+	return false
+}

pkg/config/config.go

@@ -1,6 +1,7 @@
 package config
 
 import (
+	"log"
 	"os"
 	"path/filepath"
 	"runtime"
@@ -10,9 +11,34 @@ import (
 	"gopkg.in/yaml.v3"
 )
 
+// BackendSettings contains structured backend configuration
+type BackendSettings struct {
+	Command         string            `yaml:"command"`
+	Args            []string          `yaml:"args"`
+	Environment     map[string]string `yaml:"environment,omitempty"`
+	Docker          *DockerSettings   `yaml:"docker,omitempty"`
+	ResponseHeaders map[string]string `yaml:"response_headers,omitempty"`
+}
+
+// DockerSettings contains Docker-specific configuration
+type DockerSettings struct {
+	Enabled     bool              `yaml:"enabled"`
+	Image       string            `yaml:"image"`
+	Args        []string          `yaml:"args"`
+	Environment map[string]string `yaml:"environment,omitempty"`
+}
+
+// BackendConfig contains backend executable configurations
+type BackendConfig struct {
+	LlamaCpp BackendSettings `yaml:"llama-cpp"`
+	VLLM     BackendSettings `yaml:"vllm"`
+	MLX      BackendSettings `yaml:"mlx"`
+}
+
 // AppConfig represents the configuration for llamactl
 type AppConfig struct {
 	Server     ServerConfig    `yaml:"server"`
+	Backends   BackendConfig   `yaml:"backends"`
 	Instances  InstancesConfig `yaml:"instances"`
 	Auth       AuthConfig      `yaml:"auth"`
 	Version    string          `yaml:"-"`
@@ -31,8 +57,14 @@ type ServerConfig struct {
 	// Allowed origins for CORS (e.g., "http://localhost:3000")
 	AllowedOrigins []string `yaml:"allowed_origins"`
 
+	// Allowed headers for CORS (e.g., "Accept", "Authorization", "Content-Type", "X-CSRF-Token")
+	AllowedHeaders []string `yaml:"allowed_headers"`
+
 	// Enable Swagger UI for API documentation
 	EnableSwagger bool `yaml:"enable_swagger"`
+
+	// Response headers to send with responses
+	ResponseHeaders map[string]string `yaml:"response_headers,omitempty"`
 }
 
 // InstancesConfig contains instance management configuration
@@ -55,8 +87,11 @@ type InstancesConfig struct {
 	// Maximum number of instances that can be created
 	MaxInstances int `yaml:"max_instances"`
 
-	// Path to llama-server executable
-	LlamaExecutable string `yaml:"llama_executable"`
+	// Maximum number of instances that can be running at the same time
+	MaxRunningInstances int `yaml:"max_running_instances,omitempty"`
+
+	// Enable LRU eviction for instance logs
+	EnableLRUEviction bool `yaml:"enable_lru_eviction"`
 
 	// Default auto-restart setting for new instances
 	DefaultAutoRestart bool `yaml:"default_auto_restart"`
@@ -104,16 +139,53 @@ func LoadConfig(configPath string) (AppConfig, error) {
 			Host:           "0.0.0.0",
 			Port:           8080,
 			AllowedOrigins: []string{"*"}, // Default to allow all origins
+			AllowedHeaders: []string{"*"}, // Default to allow all headers
 			EnableSwagger:  false,
 		},
+		Backends: BackendConfig{
+			LlamaCpp: BackendSettings{
+				Command:     "llama-server",
+				Args:        []string{},
+				Environment: map[string]string{},
+				Docker: &DockerSettings{
+					Enabled: false,
+					Image:   "ghcr.io/ggml-org/llama.cpp:server",
+					Args: []string{
+						"run", "--rm", "--network", "host", "--gpus", "all",
+						"-v", filepath.Join(getDefaultDataDirectory(), "llama.cpp") + ":/root/.cache/llama.cpp"},
+					Environment: map[string]string{},
+				},
+			},
+			VLLM: BackendSettings{
+				Command: "vllm",
+				Args:    []string{"serve"},
+				Docker: &DockerSettings{
+					Enabled: false,
+					Image:   "vllm/vllm-openai:latest",
+					Args: []string{
+						"run", "--rm", "--network", "host", "--gpus", "all", "--shm-size", "1g",
+						"-v", filepath.Join(getDefaultDataDirectory(), "huggingface") + ":/root/.cache/huggingface",
+					},
+					Environment: map[string]string{},
+				},
+			},
+			MLX: BackendSettings{
+				Command: "mlx_lm.server",
+				Args:    []string{},
+				// No Docker section for MLX - not supported
+			},
+		},
 		Instances: InstancesConfig{
-			PortRange:            [2]int{8000, 9000},
-			DataDir:              getDefaultDataDirectory(),
-			InstancesDir:         filepath.Join(getDefaultDataDirectory(), "instances"),
-			LogsDir:              filepath.Join(getDefaultDataDirectory(), "logs"),
+			PortRange: [2]int{8000, 9000},
+			DataDir:   getDefaultDataDirectory(),
+			// NOTE: empty strings are set as placeholder values since InstancesDir and LogsDir
+			// should be relative path to DataDir if not explicitly set.
+			InstancesDir:         "",
+			LogsDir:              "",
 			AutoCreateDirs:       true,
 			MaxInstances:         -1, // -1 means unlimited
-			LlamaExecutable:      "llama-server",
+			MaxRunningInstances:  -1, // -1 means unlimited
+			EnableLRUEviction:    true,
 			DefaultAutoRestart:   true,
 			DefaultMaxRestarts:   3,
 			DefaultRestartDelay:  5,
@@ -137,6 +209,14 @@ func LoadConfig(configPath string) (AppConfig, error) {
 	// 3. Override with environment variables
 	loadEnvVars(&cfg)
 
+	// If InstancesDir or LogsDir is not set, set it to relative path of DataDir
+	if cfg.Instances.InstancesDir == "" {
+		cfg.Instances.InstancesDir = filepath.Join(cfg.Instances.DataDir, "instances")
+	}
+	if cfg.Instances.LogsDir == "" {
+		cfg.Instances.LogsDir = filepath.Join(cfg.Instances.DataDir, "logs")
+	}
+
 	return cfg, nil
 }
 
@@ -157,6 +237,7 @@ func loadConfigFile(cfg *AppConfig, configPath string) error {
 			if err := yaml.Unmarshal(data, cfg); err != nil {
 				return err
 			}
+			log.Printf("Read config at %s", path)
 			return nil
 		}
 	}
@@ -211,9 +292,136 @@ func loadEnvVars(cfg *AppConfig) {
 			cfg.Instances.MaxInstances = m
 		}
 	}
-	if llamaExec := os.Getenv("LLAMACTL_LLAMA_EXECUTABLE"); llamaExec != "" {
-		cfg.Instances.LlamaExecutable = llamaExec
+	if maxRunning := os.Getenv("LLAMACTL_MAX_RUNNING_INSTANCES"); maxRunning != "" {
+		if m, err := strconv.Atoi(maxRunning); err == nil {
+			cfg.Instances.MaxRunningInstances = m
+		}
 	}
+	if enableLRUEviction := os.Getenv("LLAMACTL_ENABLE_LRU_EVICTION"); enableLRUEviction != "" {
+		if b, err := strconv.ParseBool(enableLRUEviction); err == nil {
+			cfg.Instances.EnableLRUEviction = b
+		}
+	}
+	// Backend config
+	// LlamaCpp backend
+	if llamaCmd := os.Getenv("LLAMACTL_LLAMACPP_COMMAND"); llamaCmd != "" {
+		cfg.Backends.LlamaCpp.Command = llamaCmd
+	}
+	if llamaArgs := os.Getenv("LLAMACTL_LLAMACPP_ARGS"); llamaArgs != "" {
+		cfg.Backends.LlamaCpp.Args = strings.Split(llamaArgs, " ")
+	}
+	if llamaEnv := os.Getenv("LLAMACTL_LLAMACPP_ENV"); llamaEnv != "" {
+		if cfg.Backends.LlamaCpp.Environment == nil {
+			cfg.Backends.LlamaCpp.Environment = make(map[string]string)
+		}
+		parseEnvVars(llamaEnv, cfg.Backends.LlamaCpp.Environment)
+	}
+	if llamaDockerEnabled := os.Getenv("LLAMACTL_LLAMACPP_DOCKER_ENABLED"); llamaDockerEnabled != "" {
+		if b, err := strconv.ParseBool(llamaDockerEnabled); err == nil {
+			if cfg.Backends.LlamaCpp.Docker == nil {
+				cfg.Backends.LlamaCpp.Docker = &DockerSettings{}
+			}
+			cfg.Backends.LlamaCpp.Docker.Enabled = b
+		}
+	}
+	if llamaDockerImage := os.Getenv("LLAMACTL_LLAMACPP_DOCKER_IMAGE"); llamaDockerImage != "" {
+		if cfg.Backends.LlamaCpp.Docker == nil {
+			cfg.Backends.LlamaCpp.Docker = &DockerSettings{}
+		}
+		cfg.Backends.LlamaCpp.Docker.Image = llamaDockerImage
+	}
+	if llamaDockerArgs := os.Getenv("LLAMACTL_LLAMACPP_DOCKER_ARGS"); llamaDockerArgs != "" {
+		if cfg.Backends.LlamaCpp.Docker == nil {
+			cfg.Backends.LlamaCpp.Docker = &DockerSettings{}
+		}
+		cfg.Backends.LlamaCpp.Docker.Args = strings.Split(llamaDockerArgs, " ")
+	}
+	if llamaDockerEnv := os.Getenv("LLAMACTL_LLAMACPP_DOCKER_ENV"); llamaDockerEnv != "" {
+		if cfg.Backends.LlamaCpp.Docker == nil {
+			cfg.Backends.LlamaCpp.Docker = &DockerSettings{}
+		}
+		if cfg.Backends.LlamaCpp.Docker.Environment == nil {
+			cfg.Backends.LlamaCpp.Docker.Environment = make(map[string]string)
+		}
+		parseEnvVars(llamaDockerEnv, cfg.Backends.LlamaCpp.Docker.Environment)
+	}
+	if llamaEnv := os.Getenv("LLAMACTL_LLAMACPP_RESPONSE_HEADERS"); llamaEnv != "" {
+		if cfg.Backends.LlamaCpp.ResponseHeaders == nil {
+			cfg.Backends.LlamaCpp.ResponseHeaders = make(map[string]string)
+		}
+		parseHeaders(llamaEnv, cfg.Backends.LlamaCpp.ResponseHeaders)
+	}
+
+	// vLLM backend
+	if vllmCmd := os.Getenv("LLAMACTL_VLLM_COMMAND"); vllmCmd != "" {
+		cfg.Backends.VLLM.Command = vllmCmd
+	}
+	if vllmArgs := os.Getenv("LLAMACTL_VLLM_ARGS"); vllmArgs != "" {
+		cfg.Backends.VLLM.Args = strings.Split(vllmArgs, " ")
+	}
+	if vllmEnv := os.Getenv("LLAMACTL_VLLM_ENV"); vllmEnv != "" {
+		if cfg.Backends.VLLM.Environment == nil {
+			cfg.Backends.VLLM.Environment = make(map[string]string)
+		}
+		parseEnvVars(vllmEnv, cfg.Backends.VLLM.Environment)
+	}
+	if vllmDockerEnabled := os.Getenv("LLAMACTL_VLLM_DOCKER_ENABLED"); vllmDockerEnabled != "" {
+		if b, err := strconv.ParseBool(vllmDockerEnabled); err == nil {
+			if cfg.Backends.VLLM.Docker == nil {
+				cfg.Backends.VLLM.Docker = &DockerSettings{}
+			}
+			cfg.Backends.VLLM.Docker.Enabled = b
+		}
+	}
+	if vllmDockerImage := os.Getenv("LLAMACTL_VLLM_DOCKER_IMAGE"); vllmDockerImage != "" {
+		if cfg.Backends.VLLM.Docker == nil {
+			cfg.Backends.VLLM.Docker = &DockerSettings{}
+		}
+		cfg.Backends.VLLM.Docker.Image = vllmDockerImage
+	}
+	if vllmDockerArgs := os.Getenv("LLAMACTL_VLLM_DOCKER_ARGS"); vllmDockerArgs != "" {
+		if cfg.Backends.VLLM.Docker == nil {
+			cfg.Backends.VLLM.Docker = &DockerSettings{}
+		}
+		cfg.Backends.VLLM.Docker.Args = strings.Split(vllmDockerArgs, " ")
+	}
+	if vllmDockerEnv := os.Getenv("LLAMACTL_VLLM_DOCKER_ENV"); vllmDockerEnv != "" {
+		if cfg.Backends.VLLM.Docker == nil {
+			cfg.Backends.VLLM.Docker = &DockerSettings{}
+		}
+		if cfg.Backends.VLLM.Docker.Environment == nil {
+			cfg.Backends.VLLM.Docker.Environment = make(map[string]string)
+		}
+		parseEnvVars(vllmDockerEnv, cfg.Backends.VLLM.Docker.Environment)
+	}
+	if llamaEnv := os.Getenv("LLAMACTL_VLLM_RESPONSE_HEADERS"); llamaEnv != "" {
+		if cfg.Backends.VLLM.ResponseHeaders == nil {
+			cfg.Backends.VLLM.ResponseHeaders = make(map[string]string)
+		}
+		parseHeaders(llamaEnv, cfg.Backends.VLLM.ResponseHeaders)
+	}
+
+	// MLX backend
+	if mlxCmd := os.Getenv("LLAMACTL_MLX_COMMAND"); mlxCmd != "" {
+		cfg.Backends.MLX.Command = mlxCmd
+	}
+	if mlxArgs := os.Getenv("LLAMACTL_MLX_ARGS"); mlxArgs != "" {
+		cfg.Backends.MLX.Args = strings.Split(mlxArgs, " ")
+	}
+	if mlxEnv := os.Getenv("LLAMACTL_MLX_ENV"); mlxEnv != "" {
+		if cfg.Backends.MLX.Environment == nil {
+			cfg.Backends.MLX.Environment = make(map[string]string)
+		}
+		parseEnvVars(mlxEnv, cfg.Backends.MLX.Environment)
+	}
+	if llamaEnv := os.Getenv("LLAMACTL_MLX_RESPONSE_HEADERS"); llamaEnv != "" {
+		if cfg.Backends.MLX.ResponseHeaders == nil {
+			cfg.Backends.MLX.ResponseHeaders = make(map[string]string)
+		}
+		parseHeaders(llamaEnv, cfg.Backends.MLX.ResponseHeaders)
+	}
+
+	// Instance defaults
 	if autoRestart := os.Getenv("LLAMACTL_DEFAULT_AUTO_RESTART"); autoRestart != "" {
 		if b, err := strconv.ParseBool(autoRestart); err == nil {
 			cfg.Instances.DefaultAutoRestart = b
@@ -286,6 +494,32 @@ func ParsePortRange(s string) [2]int {
 	return [2]int{0, 0} // Invalid format
 }
 
+// parseEnvVars parses environment variables in format "KEY1=value1,KEY2=value2"
+// and populates the provided environment map
+func parseEnvVars(envString string, envMap map[string]string) {
+	if envString == "" {
+		return
+	}
+	for _, envPair := range strings.Split(envString, ",") {
+		if parts := strings.SplitN(strings.TrimSpace(envPair), "=", 2); len(parts) == 2 {
+			envMap[parts[0]] = parts[1]
+		}
+	}
+}
+
+// parseHeaders parses HTTP headers in format "KEY1=value1;KEY2=value2"
+// and populates the provided environment map
+func parseHeaders(envString string, envMap map[string]string) {
+	if envString == "" {
+		return
+	}
+	for _, envPair := range strings.Split(envString, ";") {
+		if parts := strings.SplitN(strings.TrimSpace(envPair), "=", 2); len(parts) == 2 {
+			envMap[parts[0]] = parts[1]
+		}
+	}
+}
+
 // getDefaultDataDirectory returns platform-specific default data directory
 func getDefaultDataDirectory() string {
 	switch runtime.GOOS {
@@ -318,6 +552,10 @@ func getDefaultDataDirectory() string {
 // getDefaultConfigLocations returns platform-specific config file locations
 func getDefaultConfigLocations() []string {
 	var locations []string
+	// Use ./llamactl.yaml and ./config.yaml as the default config file
+	locations = append(locations, "llamactl.yaml")
+	locations = append(locations, "config.yaml")
+
 	homeDir, _ := os.UserHomeDir()
 
 	switch runtime.GOOS {
@@ -347,3 +585,17 @@ func getDefaultConfigLocations() []string {
 
 	return locations
 }
+
+// GetBackendSettings resolves backend settings
+func (bc *BackendConfig) GetBackendSettings(backendType string) BackendSettings {
+	switch backendType {
+	case "llama-cpp":
+		return bc.LlamaCpp
+	case "vllm":
+		return bc.VLLM
+	case "mlx":
+		return bc.MLX
+	default:
+		return BackendSettings{}
+	}
+}

pkg/config/config_test.go

@@ -42,9 +42,6 @@ func TestLoadConfig_Defaults(t *testing.T) {
 	if cfg.Instances.MaxInstances != -1 {
 		t.Errorf("Expected default max instances -1, got %d", cfg.Instances.MaxInstances)
 	}
-	if cfg.Instances.LlamaExecutable != "llama-server" {
-		t.Errorf("Expected default executable 'llama-server', got %q", cfg.Instances.LlamaExecutable)
-	}
 	if !cfg.Instances.DefaultAutoRestart {
 		t.Error("Expected default auto restart to be true")
 	}
@@ -101,9 +98,6 @@ instances:
 	if cfg.Instances.MaxInstances != 5 {
 		t.Errorf("Expected max instances 5, got %d", cfg.Instances.MaxInstances)
 	}
-	if cfg.Instances.LlamaExecutable != "/usr/bin/llama-server" {
-		t.Errorf("Expected executable '/usr/bin/llama-server', got %q", cfg.Instances.LlamaExecutable)
-	}
 	if cfg.Instances.DefaultAutoRestart {
 		t.Error("Expected auto restart to be false")
 	}
@@ -123,7 +117,6 @@ func TestLoadConfig_EnvironmentOverrides(t *testing.T) {
 		"LLAMACTL_INSTANCE_PORT_RANGE":   "5000-6000",
 		"LLAMACTL_LOGS_DIR":              "/env/logs",
 		"LLAMACTL_MAX_INSTANCES":         "20",
-		"LLAMACTL_LLAMA_EXECUTABLE":      "/env/llama-server",
 		"LLAMACTL_DEFAULT_AUTO_RESTART":  "false",
 		"LLAMACTL_DEFAULT_MAX_RESTARTS":  "7",
 		"LLAMACTL_DEFAULT_RESTART_DELAY": "15",
@@ -156,8 +149,8 @@ func TestLoadConfig_EnvironmentOverrides(t *testing.T) {
 	if cfg.Instances.MaxInstances != 20 {
 		t.Errorf("Expected max instances 20, got %d", cfg.Instances.MaxInstances)
 	}
-	if cfg.Instances.LlamaExecutable != "/env/llama-server" {
-		t.Errorf("Expected executable '/env/llama-server', got %q", cfg.Instances.LlamaExecutable)
+	if cfg.Backends.LlamaCpp.Command != "llama-server" {
+		t.Errorf("Expected default llama command 'llama-server', got %q", cfg.Backends.LlamaCpp.Command)
 	}
 	if cfg.Instances.DefaultAutoRestart {
 		t.Error("Expected auto restart to be false")
@@ -355,3 +348,165 @@ server:
 		t.Errorf("Expected default max instances -1, got %d", cfg.Instances.MaxInstances)
 	}
 }
+
+func TestGetBackendSettings_NewStructuredConfig(t *testing.T) {
+	bc := &config.BackendConfig{
+		LlamaCpp: config.BackendSettings{
+			Command: "custom-llama",
+			Args:    []string{"--verbose"},
+			Docker: &config.DockerSettings{
+				Enabled:     true,
+				Image:       "custom-llama:latest",
+				Args:        []string{"--gpus", "all"},
+				Environment: map[string]string{"CUDA_VISIBLE_DEVICES": "1"},
+			},
+		},
+		VLLM: config.BackendSettings{
+			Command: "custom-vllm",
+			Args:    []string{"serve", "--debug"},
+		},
+		MLX: config.BackendSettings{
+			Command: "custom-mlx",
+			Args:    []string{},
+		},
+	}
+
+	// Test llama-cpp with Docker
+	settings := bc.GetBackendSettings("llama-cpp")
+	if settings.Command != "custom-llama" {
+		t.Errorf("Expected command 'custom-llama', got %q", settings.Command)
+	}
+	if len(settings.Args) != 1 || settings.Args[0] != "--verbose" {
+		t.Errorf("Expected args ['--verbose'], got %v", settings.Args)
+	}
+	if settings.Docker == nil || !settings.Docker.Enabled {
+		t.Error("Expected Docker to be enabled")
+	}
+	if settings.Docker.Image != "custom-llama:latest" {
+		t.Errorf("Expected Docker image 'custom-llama:latest', got %q", settings.Docker.Image)
+	}
+
+	// Test vLLM without Docker
+	settings = bc.GetBackendSettings("vllm")
+	if settings.Command != "custom-vllm" {
+		t.Errorf("Expected command 'custom-vllm', got %q", settings.Command)
+	}
+	if len(settings.Args) != 2 || settings.Args[0] != "serve" || settings.Args[1] != "--debug" {
+		t.Errorf("Expected args ['serve', '--debug'], got %v", settings.Args)
+	}
+	if settings.Docker != nil && settings.Docker.Enabled {
+		t.Error("Expected Docker to be disabled or nil")
+	}
+
+	// Test MLX
+	settings = bc.GetBackendSettings("mlx")
+	if settings.Command != "custom-mlx" {
+		t.Errorf("Expected command 'custom-mlx', got %q", settings.Command)
+	}
+}
+
+func TestGetBackendSettings_EmptyConfig(t *testing.T) {
+	bc := &config.BackendConfig{}
+
+	// Test empty llama-cpp
+	settings := bc.GetBackendSettings("llama-cpp")
+	if settings.Command != "" {
+		t.Errorf("Expected empty command, got %q", settings.Command)
+	}
+
+	// Test empty vLLM
+	settings = bc.GetBackendSettings("vllm")
+	if settings.Command != "" {
+		t.Errorf("Expected empty command, got %q", settings.Command)
+	}
+
+	// Test empty MLX
+	settings = bc.GetBackendSettings("mlx")
+	if settings.Command != "" {
+		t.Errorf("Expected empty command, got %q", settings.Command)
+	}
+}
+
+func TestLoadConfig_BackendEnvironmentVariables(t *testing.T) {
+	// Test that backend environment variables work correctly
+	envVars := map[string]string{
+		"LLAMACTL_LLAMACPP_COMMAND":        "env-llama",
+		"LLAMACTL_LLAMACPP_ARGS":           "--verbose --threads 4",
+		"LLAMACTL_LLAMACPP_DOCKER_ENABLED": "true",
+		"LLAMACTL_LLAMACPP_DOCKER_IMAGE":   "env-llama:latest",
+		"LLAMACTL_LLAMACPP_DOCKER_ARGS":    "run --rm --network host --gpus all",
+		"LLAMACTL_LLAMACPP_DOCKER_ENV":     "CUDA_VISIBLE_DEVICES=0,OMP_NUM_THREADS=4",
+		"LLAMACTL_VLLM_COMMAND":            "env-vllm",
+		"LLAMACTL_VLLM_DOCKER_ENABLED":     "false",
+		"LLAMACTL_VLLM_DOCKER_IMAGE":       "env-vllm:latest",
+		"LLAMACTL_VLLM_DOCKER_ENV":         "PYTORCH_CUDA_ALLOC_CONF=max_split_size_mb:512,CUDA_VISIBLE_DEVICES=1",
+		"LLAMACTL_MLX_COMMAND":             "env-mlx",
+	}
+
+	// Set env vars and ensure cleanup
+	for key, value := range envVars {
+		os.Setenv(key, value)
+		defer os.Unsetenv(key)
+	}
+
+	cfg, err := config.LoadConfig("nonexistent-file.yaml")
+	if err != nil {
+		t.Fatalf("LoadConfig failed: %v", err)
+	}
+
+	// Verify llama-cpp environment overrides
+	if cfg.Backends.LlamaCpp.Command != "env-llama" {
+		t.Errorf("Expected llama command 'env-llama', got %q", cfg.Backends.LlamaCpp.Command)
+	}
+	expectedArgs := []string{"--verbose", "--threads", "4"}
+	if len(cfg.Backends.LlamaCpp.Args) != len(expectedArgs) {
+		t.Errorf("Expected llama args %v, got %v", expectedArgs, cfg.Backends.LlamaCpp.Args)
+	}
+	if !cfg.Backends.LlamaCpp.Docker.Enabled {
+		t.Error("Expected llama Docker to be enabled")
+	}
+	if cfg.Backends.LlamaCpp.Docker.Image != "env-llama:latest" {
+		t.Errorf("Expected llama Docker image 'env-llama:latest', got %q", cfg.Backends.LlamaCpp.Docker.Image)
+	}
+	expectedDockerArgs := []string{"run", "--rm", "--network", "host", "--gpus", "all"}
+	if len(cfg.Backends.LlamaCpp.Docker.Args) != len(expectedDockerArgs) {
+		t.Errorf("Expected llama Docker args %v, got %v", expectedDockerArgs, cfg.Backends.LlamaCpp.Docker.Args)
+	}
+	if cfg.Backends.LlamaCpp.Docker.Environment["CUDA_VISIBLE_DEVICES"] != "0" {
+		t.Errorf("Expected CUDA_VISIBLE_DEVICES=0, got %q", cfg.Backends.LlamaCpp.Docker.Environment["CUDA_VISIBLE_DEVICES"])
+	}
+	if cfg.Backends.LlamaCpp.Docker.Environment["OMP_NUM_THREADS"] != "4" {
+		t.Errorf("Expected OMP_NUM_THREADS=4, got %q", cfg.Backends.LlamaCpp.Docker.Environment["OMP_NUM_THREADS"])
+	}
+
+	// Verify vLLM environment overrides
+	if cfg.Backends.VLLM.Command != "env-vllm" {
+		t.Errorf("Expected vLLM command 'env-vllm', got %q", cfg.Backends.VLLM.Command)
+	}
+	if cfg.Backends.VLLM.Docker.Enabled {
+		t.Error("Expected vLLM Docker to be disabled")
+	}
+	if cfg.Backends.VLLM.Docker.Environment["PYTORCH_CUDA_ALLOC_CONF"] != "max_split_size_mb:512" {
+		t.Errorf("Expected PYTORCH_CUDA_ALLOC_CONF=max_split_size_mb:512, got %q", cfg.Backends.VLLM.Docker.Environment["PYTORCH_CUDA_ALLOC_CONF"])
+	}
+
+	// Verify MLX environment overrides
+	if cfg.Backends.MLX.Command != "env-mlx" {
+		t.Errorf("Expected MLX command 'env-mlx', got %q", cfg.Backends.MLX.Command)
+	}
+}
+
+func TestGetBackendSettings_InvalidBackendType(t *testing.T) {
+	bc := &config.BackendConfig{
+		LlamaCpp: config.BackendSettings{
+			Command: "llama-server",
+			Args:    []string{},
+		},
+	}
+
+	// Test invalid backend type returns empty settings
+	settings := bc.GetBackendSettings("invalid-backend")
+	if settings.Command != "" {
+		t.Errorf("Expected empty command for invalid backend, got %q", settings.Command)
+	}
+}

pkg/instance/instance.go

@@ -5,7 +5,7 @@ import (
 	"encoding/json"
 	"fmt"
 	"io"
-	"llamactl/pkg/backends/llamacpp"
+	"llamactl/pkg/backends"
 	"llamactl/pkg/config"
 	"log"
 	"net/http"
@@ -29,60 +29,16 @@ func (realTimeProvider) Now() time.Time {
 	return time.Now()
 }
 
-type CreateInstanceOptions struct {
-	// Auto restart
-	AutoRestart  *bool `json:"auto_restart,omitempty"`
-	MaxRestarts  *int  `json:"max_restarts,omitempty"`
-	RestartDelay *int  `json:"restart_delay,omitempty"`
-	// On demand start
-	OnDemandStart *bool `json:"on_demand_start,omitempty"`
-	// Idle timeout
-	IdleTimeout *int `json:"idle_timeout,omitempty"`
-	// LlamaServerOptions contains the options for the llama server
-	llamacpp.LlamaServerOptions `json:",inline"`
-}
-
-// UnmarshalJSON implements custom JSON unmarshaling for CreateInstanceOptions
-// This is needed because the embedded LlamaServerOptions has its own UnmarshalJSON
-// which can interfere with proper unmarshaling of the pointer fields
-func (c *CreateInstanceOptions) UnmarshalJSON(data []byte) error {
-	// First, unmarshal into a temporary struct without the embedded type
-	type tempCreateOptions struct {
-		AutoRestart   *bool `json:"auto_restart,omitempty"`
-		MaxRestarts   *int  `json:"max_restarts,omitempty"`
-		RestartDelay  *int  `json:"restart_delay,omitempty"`
-		OnDemandStart *bool `json:"on_demand_start,omitempty"`
-		IdleTimeout   *int  `json:"idle_timeout,omitempty"`
-	}
-
-	var temp tempCreateOptions
-	if err := json.Unmarshal(data, &temp); err != nil {
-		return err
-	}
-
-	// Copy the pointer fields
-	c.AutoRestart = temp.AutoRestart
-	c.MaxRestarts = temp.MaxRestarts
-	c.RestartDelay = temp.RestartDelay
-	c.OnDemandStart = temp.OnDemandStart
-	c.IdleTimeout = temp.IdleTimeout
-
-	// Now unmarshal the embedded LlamaServerOptions
-	if err := json.Unmarshal(data, &c.LlamaServerOptions); err != nil {
-		return err
-	}
-
-	return nil
-}
-
 // Process represents a running instance of the llama server
 type Process struct {
-	Name           string                 `json:"name"`
-	options        *CreateInstanceOptions `json:"-"`
-	globalSettings *config.InstancesConfig
+	Name                   string                 `json:"name"`
+	options                *CreateInstanceOptions `json:"-"`
+	globalInstanceSettings *config.InstancesConfig
+	globalBackendSettings  *config.BackendConfig
 
 	// Status
-	Running bool `json:"running"`
+	Status         InstanceStatus `json:"status"`
+	onStatusChange func(oldStatus, newStatus InstanceStatus)
 
 	// Creation time
 	Created int64 `json:"created,omitempty"` // Unix timestamp when the instance was created
@@ -109,105 +65,24 @@ type Process struct {
 	timeProvider    TimeProvider `json:"-"` // Time provider for testing
 }
 
-// validateAndCopyOptions validates and creates a deep copy of the provided options
-// It applies validation rules and returns a safe copy
-func validateAndCopyOptions(name string, options *CreateInstanceOptions) *CreateInstanceOptions {
-	optionsCopy := &CreateInstanceOptions{}
-
-	if options != nil {
-		// Copy the embedded LlamaServerOptions
-		optionsCopy.LlamaServerOptions = options.LlamaServerOptions
-
-		// Copy and validate pointer fields
-		if options.AutoRestart != nil {
-			autoRestart := *options.AutoRestart
-			optionsCopy.AutoRestart = &autoRestart
-		}
-
-		if options.MaxRestarts != nil {
-			maxRestarts := *options.MaxRestarts
-			if maxRestarts < 0 {
-				log.Printf("Instance %s MaxRestarts value (%d) cannot be negative, setting to 0", name, maxRestarts)
-				maxRestarts = 0
-			}
-			optionsCopy.MaxRestarts = &maxRestarts
-		}
-
-		if options.RestartDelay != nil {
-			restartDelay := *options.RestartDelay
-			if restartDelay < 0 {
-				log.Printf("Instance %s RestartDelay value (%d) cannot be negative, setting to 0 seconds", name, restartDelay)
-				restartDelay = 0
-			}
-			optionsCopy.RestartDelay = &restartDelay
-		}
-
-		if options.OnDemandStart != nil {
-			onDemandStart := *options.OnDemandStart
-			optionsCopy.OnDemandStart = &onDemandStart
-		}
-
-		if options.IdleTimeout != nil {
-			idleTimeout := *options.IdleTimeout
-			if idleTimeout < 0 {
-				log.Printf("Instance %s IdleTimeout value (%d) cannot be negative, setting to 0 minutes", name, idleTimeout)
-				idleTimeout = 0
-			}
-			optionsCopy.IdleTimeout = &idleTimeout
-		}
-	}
-
-	return optionsCopy
-}
-
-// applyDefaultOptions applies default values from global settings to any nil options
-func applyDefaultOptions(options *CreateInstanceOptions, globalSettings *config.InstancesConfig) {
-	if globalSettings == nil {
-		return
-	}
-
-	if options.AutoRestart == nil {
-		defaultAutoRestart := globalSettings.DefaultAutoRestart
-		options.AutoRestart = &defaultAutoRestart
-	}
-
-	if options.MaxRestarts == nil {
-		defaultMaxRestarts := globalSettings.DefaultMaxRestarts
-		options.MaxRestarts = &defaultMaxRestarts
-	}
-
-	if options.RestartDelay == nil {
-		defaultRestartDelay := globalSettings.DefaultRestartDelay
-		options.RestartDelay = &defaultRestartDelay
-	}
-
-	if options.OnDemandStart == nil {
-		defaultOnDemandStart := globalSettings.DefaultOnDemandStart
-		options.OnDemandStart = &defaultOnDemandStart
-	}
-
-	if options.IdleTimeout == nil {
-		defaultIdleTimeout := 0
-		options.IdleTimeout = &defaultIdleTimeout
-	}
-}
-
 // NewInstance creates a new instance with the given name, log path, and options
-func NewInstance(name string, globalSettings *config.InstancesConfig, options *CreateInstanceOptions) *Process {
+func NewInstance(name string, globalBackendSettings *config.BackendConfig, globalInstanceSettings *config.InstancesConfig, options *CreateInstanceOptions, onStatusChange func(oldStatus, newStatus InstanceStatus)) *Process {
 	// Validate and copy options
-	optionsCopy := validateAndCopyOptions(name, options)
-	// Apply defaults
-	applyDefaultOptions(optionsCopy, globalSettings)
+	options.ValidateAndApplyDefaults(name, globalInstanceSettings)
+
 	// Create the instance logger
-	logger := NewInstanceLogger(name, globalSettings.LogsDir)
+	logger := NewInstanceLogger(name, globalInstanceSettings.LogsDir)
 
 	return &Process{
-		Name:           name,
-		options:        optionsCopy,
-		globalSettings: globalSettings,
-		logger:         logger,
-		timeProvider:   realTimeProvider{},
-		Created:        time.Now().Unix(),
+		Name:                   name,
+		options:                options,
+		globalInstanceSettings: globalInstanceSettings,
+		globalBackendSettings:  globalBackendSettings,
+		logger:                 logger,
+		timeProvider:           realTimeProvider{},
+		Created:                time.Now().Unix(),
+		Status:                 Stopped,
+		onStatusChange:         onStatusChange,
 	}
 }
 
@@ -217,6 +92,50 @@ func (i *Process) GetOptions() *CreateInstanceOptions {
 	return i.options
 }
 
+func (i *Process) GetPort() int {
+	i.mu.RLock()
+	defer i.mu.RUnlock()
+	if i.options != nil {
+		switch i.options.BackendType {
+		case backends.BackendTypeLlamaCpp:
+			if i.options.LlamaServerOptions != nil {
+				return i.options.LlamaServerOptions.Port
+			}
+		case backends.BackendTypeMlxLm:
+			if i.options.MlxServerOptions != nil {
+				return i.options.MlxServerOptions.Port
+			}
+		case backends.BackendTypeVllm:
+			if i.options.VllmServerOptions != nil {
+				return i.options.VllmServerOptions.Port
+			}
+		}
+	}
+	return 0
+}
+
+func (i *Process) GetHost() string {
+	i.mu.RLock()
+	defer i.mu.RUnlock()
+	if i.options != nil {
+		switch i.options.BackendType {
+		case backends.BackendTypeLlamaCpp:
+			if i.options.LlamaServerOptions != nil {
+				return i.options.LlamaServerOptions.Host
+			}
+		case backends.BackendTypeMlxLm:
+			if i.options.MlxServerOptions != nil {
+				return i.options.MlxServerOptions.Host
+			}
+		case backends.BackendTypeVllm:
+			if i.options.VllmServerOptions != nil {
+				return i.options.VllmServerOptions.Host
+			}
+		}
+	}
+	return ""
+}
+
 func (i *Process) SetOptions(options *CreateInstanceOptions) {
 	i.mu.Lock()
 	defer i.mu.Unlock()
@@ -226,11 +145,10 @@ func (i *Process) SetOptions(options *CreateInstanceOptions) {
 		return
 	}
 
-	// Validate and copy options and apply defaults
-	optionsCopy := validateAndCopyOptions(i.Name, options)
-	applyDefaultOptions(optionsCopy, i.globalSettings)
+	// Validate and copy options
+	options.ValidateAndApplyDefaults(i.Name, i.globalInstanceSettings)
 
-	i.options = optionsCopy
+	i.options = options
 	// Clear the proxy so it gets recreated with new options
 	i.proxy = nil
 }
@@ -253,13 +171,42 @@ func (i *Process) GetProxy() (*httputil.ReverseProxy, error) {
 		return nil, fmt.Errorf("instance %s has no options set", i.Name)
 	}
 
-	targetURL, err := url.Parse(fmt.Sprintf("http://%s:%d", i.options.Host, i.options.Port))
+	var host string
+	var port int
+	switch i.options.BackendType {
+	case backends.BackendTypeLlamaCpp:
+		if i.options.LlamaServerOptions != nil {
+			host = i.options.LlamaServerOptions.Host
+			port = i.options.LlamaServerOptions.Port
+		}
+	case backends.BackendTypeMlxLm:
+		if i.options.MlxServerOptions != nil {
+			host = i.options.MlxServerOptions.Host
+			port = i.options.MlxServerOptions.Port
+		}
+	case backends.BackendTypeVllm:
+		if i.options.VllmServerOptions != nil {
+			host = i.options.VllmServerOptions.Host
+			port = i.options.VllmServerOptions.Port
+		}
+	}
+
+	targetURL, err := url.Parse(fmt.Sprintf("http://%s:%d", host, port))
 	if err != nil {
 		return nil, fmt.Errorf("failed to parse target URL for instance %s: %w", i.Name, err)
 	}
 
 	proxy := httputil.NewSingleHostReverseProxy(targetURL)
 
+	var responseHeaders map[string]string
+	switch i.options.BackendType {
+	case backends.BackendTypeLlamaCpp:
+		responseHeaders = i.globalBackendSettings.LlamaCpp.ResponseHeaders
+	case backends.BackendTypeVllm:
+		responseHeaders = i.globalBackendSettings.VLLM.ResponseHeaders
+	case backends.BackendTypeMlxLm:
+		responseHeaders = i.globalBackendSettings.MLX.ResponseHeaders
+	}
 	proxy.ModifyResponse = func(resp *http.Response) error {
 		// Remove CORS headers from llama-server response to avoid conflicts
 		// llamactl will add its own CORS headers
@@ -269,6 +216,10 @@ func (i *Process) GetProxy() (*httputil.ReverseProxy, error) {
 		resp.Header.Del("Access-Control-Allow-Credentials")
 		resp.Header.Del("Access-Control-Max-Age")
 		resp.Header.Del("Access-Control-Expose-Headers")
+
+		for key, value := range responseHeaders {
+			resp.Header.Set(key, value)
+		}
 		return nil
 	}
 
@@ -283,44 +234,55 @@ func (i *Process) MarshalJSON() ([]byte, error) {
 	i.mu.RLock()
 	defer i.mu.RUnlock()
 
-	// Create a temporary struct with exported fields for JSON marshalling
-	temp := struct {
-		Name    string                 `json:"name"`
-		Options *CreateInstanceOptions `json:"options,omitempty"`
-		Running bool                   `json:"running"`
-		Created int64                  `json:"created,omitempty"`
-	}{
-		Name:    i.Name,
-		Options: i.options,
-		Running: i.Running,
-		Created: i.Created,
+	// Determine if docker is enabled for this instance's backend
+	var dockerEnabled bool
+	if i.options != nil {
+		switch i.options.BackendType {
+		case backends.BackendTypeLlamaCpp:
+			if i.globalBackendSettings != nil && i.globalBackendSettings.LlamaCpp.Docker != nil && i.globalBackendSettings.LlamaCpp.Docker.Enabled {
+				dockerEnabled = true
+			}
+		case backends.BackendTypeVllm:
+			if i.globalBackendSettings != nil && i.globalBackendSettings.VLLM.Docker != nil && i.globalBackendSettings.VLLM.Docker.Enabled {
+				dockerEnabled = true
+			}
+		case backends.BackendTypeMlxLm:
+			// MLX does not support docker currently
+		}
 	}
 
-	return json.Marshal(temp)
+	// Use anonymous struct to avoid recursion
+	type Alias Process
+	return json.Marshal(&struct {
+		*Alias
+		Options       *CreateInstanceOptions `json:"options,omitempty"`
+		DockerEnabled bool                   `json:"docker_enabled,omitempty"`
+	}{
+		Alias:         (*Alias)(i),
+		Options:       i.options,
+		DockerEnabled: dockerEnabled,
+	})
 }
 
 // UnmarshalJSON implements json.Unmarshaler for Instance
 func (i *Process) UnmarshalJSON(data []byte) error {
-	// Create a temporary struct for unmarshalling
-	temp := struct {
-		Name    string                 `json:"name"`
+	// Use anonymous struct to avoid recursion
+	type Alias Process
+	aux := &struct {
+		*Alias
 		Options *CreateInstanceOptions `json:"options,omitempty"`
-		Running bool                   `json:"running"`
-		Created int64                  `json:"created,omitempty"`
-	}{}
+	}{
+		Alias: (*Alias)(i),
+	}
 
-	if err := json.Unmarshal(data, &temp); err != nil {
+	if err := json.Unmarshal(data, aux); err != nil {
 		return err
 	}
 
-	// Set the fields
-	i.Name = temp.Name
-	i.Running = temp.Running
-	i.Created = temp.Created
-
-	// Handle options with validation but no defaults
-	if temp.Options != nil {
-		i.options = validateAndCopyOptions(i.Name, temp.Options)
+	// Handle options with validation and defaults
+	if aux.Options != nil {
+		aux.Options.ValidateAndApplyDefaults(i.Name, i.globalInstanceSettings)
+		i.options = aux.Options
 	}
 
 	return nil

pkg/instance/instance_test.go

@@ -2,6 +2,7 @@ package instance_test
 
 import (
 	"encoding/json"
+	"llamactl/pkg/backends"
 	"llamactl/pkg/backends/llamacpp"
 	"llamactl/pkg/config"
 	"llamactl/pkg/instance"
@@ -10,6 +11,21 @@ import (
 )
 
 func TestNewInstance(t *testing.T) {
+	backendConfig := &config.BackendConfig{
+		LlamaCpp: config.BackendSettings{
+			Command: "llama-server",
+			Args:    []string{},
+		},
+		MLX: config.BackendSettings{
+			Command: "mlx_lm.server",
+			Args:    []string{},
+		},
+		VLLM: config.BackendSettings{
+			Command: "vllm",
+			Args:    []string{"serve"},
+		},
+	}
+
 	globalSettings := &config.InstancesConfig{
 		LogsDir:             "/tmp/test",
 		DefaultAutoRestart:  true,
@@ -18,28 +34,32 @@ func TestNewInstance(t *testing.T) {
 	}
 
 	options := &instance.CreateInstanceOptions{
-		LlamaServerOptions: llamacpp.LlamaServerOptions{
+		BackendType: backends.BackendTypeLlamaCpp,
+		LlamaServerOptions: &llamacpp.LlamaServerOptions{
 			Model: "/path/to/model.gguf",
 			Port:  8080,
 		},
 	}
 
-	instance := instance.NewInstance("test-instance", globalSettings, options)
+	// Mock onStatusChange function
+	mockOnStatusChange := func(oldStatus, newStatus instance.InstanceStatus) {}
 
-	if instance.Name != "test-instance" {
-		t.Errorf("Expected name 'test-instance', got %q", instance.Name)
+	inst := instance.NewInstance("test-instance", backendConfig, globalSettings, options, mockOnStatusChange)
+
+	if inst.Name != "test-instance" {
+		t.Errorf("Expected name 'test-instance', got %q", inst.Name)
 	}
-	if instance.Running {
+	if inst.IsRunning() {
 		t.Error("New instance should not be running")
 	}
 
 	// Check that options were properly set with defaults applied
-	opts := instance.GetOptions()
-	if opts.Model != "/path/to/model.gguf" {
-		t.Errorf("Expected model '/path/to/model.gguf', got %q", opts.Model)
+	opts := inst.GetOptions()
+	if opts.LlamaServerOptions.Model != "/path/to/model.gguf" {
+		t.Errorf("Expected model '/path/to/model.gguf', got %q", opts.LlamaServerOptions.Model)
 	}
-	if opts.Port != 8080 {
-		t.Errorf("Expected port 8080, got %d", opts.Port)
+	if inst.GetPort() != 8080 {
+		t.Errorf("Expected port 8080, got %d", inst.GetPort())
 	}
 
 	// Check that defaults were applied
@@ -55,6 +75,21 @@ func TestNewInstance(t *testing.T) {
 }
 
 func TestNewInstance_WithRestartOptions(t *testing.T) {
+	backendConfig := &config.BackendConfig{
+		LlamaCpp: config.BackendSettings{
+			Command: "llama-server",
+			Args:    []string{},
+		},
+		MLX: config.BackendSettings{
+			Command: "mlx_lm.server",
+			Args:    []string{},
+		},
+		VLLM: config.BackendSettings{
+			Command: "vllm",
+			Args:    []string{"serve"},
+		},
+	}
+
 	globalSettings := &config.InstancesConfig{
 		LogsDir:             "/tmp/test",
 		DefaultAutoRestart:  true,
@@ -71,12 +106,16 @@ func TestNewInstance_WithRestartOptions(t *testing.T) {
 		AutoRestart:  &autoRestart,
 		MaxRestarts:  &maxRestarts,
 		RestartDelay: &restartDelay,
-		LlamaServerOptions: llamacpp.LlamaServerOptions{
+		BackendType:  backends.BackendTypeLlamaCpp,
+		LlamaServerOptions: &llamacpp.LlamaServerOptions{
 			Model: "/path/to/model.gguf",
 		},
 	}
 
-	instance := instance.NewInstance("test-instance", globalSettings, options)
+	// Mock onStatusChange function
+	mockOnStatusChange := func(oldStatus, newStatus instance.InstanceStatus) {}
+
+	instance := instance.NewInstance("test-instance", backendConfig, globalSettings, options, mockOnStatusChange)
 	opts := instance.GetOptions()
 
 	// Check that explicit values override defaults
@@ -92,6 +131,21 @@ func TestNewInstance_WithRestartOptions(t *testing.T) {
 }
 
 func TestSetOptions(t *testing.T) {
+	backendConfig := &config.BackendConfig{
+		LlamaCpp: config.BackendSettings{
+			Command: "llama-server",
+			Args:    []string{},
+		},
+		MLX: config.BackendSettings{
+			Command: "mlx_lm.server",
+			Args:    []string{},
+		},
+		VLLM: config.BackendSettings{
+			Command: "vllm",
+			Args:    []string{"serve"},
+		},
+	}
+
 	globalSettings := &config.InstancesConfig{
 		LogsDir:             "/tmp/test",
 		DefaultAutoRestart:  true,
@@ -100,17 +154,22 @@ func TestSetOptions(t *testing.T) {
 	}
 
 	initialOptions := &instance.CreateInstanceOptions{
-		LlamaServerOptions: llamacpp.LlamaServerOptions{
+		BackendType: backends.BackendTypeLlamaCpp,
+		LlamaServerOptions: &llamacpp.LlamaServerOptions{
 			Model: "/path/to/model.gguf",
 			Port:  8080,
 		},
 	}
 
-	inst := instance.NewInstance("test-instance", globalSettings, initialOptions)
+	// Mock onStatusChange function
+	mockOnStatusChange := func(oldStatus, newStatus instance.InstanceStatus) {}
+
+	inst := instance.NewInstance("test-instance", backendConfig, globalSettings, initialOptions, mockOnStatusChange)
 
 	// Update options
 	newOptions := &instance.CreateInstanceOptions{
-		LlamaServerOptions: llamacpp.LlamaServerOptions{
+		BackendType: backends.BackendTypeLlamaCpp,
+		LlamaServerOptions: &llamacpp.LlamaServerOptions{
 			Model: "/path/to/new-model.gguf",
 			Port:  8081,
 		},
@@ -119,11 +178,11 @@ func TestSetOptions(t *testing.T) {
 	inst.SetOptions(newOptions)
 	opts := inst.GetOptions()
 
-	if opts.Model != "/path/to/new-model.gguf" {
-		t.Errorf("Expected updated model '/path/to/new-model.gguf', got %q", opts.Model)
+	if opts.LlamaServerOptions.Model != "/path/to/new-model.gguf" {
+		t.Errorf("Expected updated model '/path/to/new-model.gguf', got %q", opts.LlamaServerOptions.Model)
 	}
-	if opts.Port != 8081 {
-		t.Errorf("Expected updated port 8081, got %d", opts.Port)
+	if inst.GetPort() != 8081 {
+		t.Errorf("Expected updated port 8081, got %d", inst.GetPort())
 	}
 
 	// Check that defaults are still applied
@@ -133,18 +192,37 @@ func TestSetOptions(t *testing.T) {
 }
 
 func TestGetProxy(t *testing.T) {
+	backendConfig := &config.BackendConfig{
+		LlamaCpp: config.BackendSettings{
+			Command: "llama-server",
+			Args:    []string{},
+		},
+		MLX: config.BackendSettings{
+			Command: "mlx_lm.server",
+			Args:    []string{},
+		},
+		VLLM: config.BackendSettings{
+			Command: "vllm",
+			Args:    []string{"serve"},
+		},
+	}
+
 	globalSettings := &config.InstancesConfig{
 		LogsDir: "/tmp/test",
 	}
 
 	options := &instance.CreateInstanceOptions{
-		LlamaServerOptions: llamacpp.LlamaServerOptions{
+		BackendType: backends.BackendTypeLlamaCpp,
+		LlamaServerOptions: &llamacpp.LlamaServerOptions{
 			Host: "localhost",
 			Port: 8080,
 		},
 	}
 
-	inst := instance.NewInstance("test-instance", globalSettings, options)
+	// Mock onStatusChange function
+	mockOnStatusChange := func(oldStatus, newStatus instance.InstanceStatus) {}
+
+	inst := instance.NewInstance("test-instance", backendConfig, globalSettings, options, mockOnStatusChange)
 
 	// Get proxy for the first time
 	proxy1, err := inst.GetProxy()
@@ -166,6 +244,21 @@ func TestGetProxy(t *testing.T) {
 }
 
 func TestMarshalJSON(t *testing.T) {
+	backendConfig := &config.BackendConfig{
+		LlamaCpp: config.BackendSettings{
+			Command: "llama-server",
+			Args:    []string{},
+		},
+		MLX: config.BackendSettings{
+			Command: "mlx_lm.server",
+			Args:    []string{},
+		},
+		VLLM: config.BackendSettings{
+			Command: "vllm",
+			Args:    []string{"serve"},
+		},
+	}
+
 	globalSettings := &config.InstancesConfig{
 		LogsDir:             "/tmp/test",
 		DefaultAutoRestart:  true,
@@ -174,13 +267,17 @@ func TestMarshalJSON(t *testing.T) {
 	}
 
 	options := &instance.CreateInstanceOptions{
-		LlamaServerOptions: llamacpp.LlamaServerOptions{
+		BackendType: backends.BackendTypeLlamaCpp,
+		LlamaServerOptions: &llamacpp.LlamaServerOptions{
 			Model: "/path/to/model.gguf",
 			Port:  8080,
 		},
 	}
 
-	instance := instance.NewInstance("test-instance", globalSettings, options)
+	// Mock onStatusChange function
+	mockOnStatusChange := func(oldStatus, newStatus instance.InstanceStatus) {}
+
+	instance := instance.NewInstance("test-instance", backendConfig, globalSettings, options, mockOnStatusChange)
 
 	data, err := json.Marshal(instance)
 	if err != nil {
@@ -188,7 +285,7 @@ func TestMarshalJSON(t *testing.T) {
 	}
 
 	// Check that JSON contains expected fields
-	var result map[string]interface{}
+	var result map[string]any
 	err = json.Unmarshal(data, &result)
 	if err != nil {
 		t.Fatalf("JSON unmarshal failed: %v", err)
@@ -197,8 +294,8 @@ func TestMarshalJSON(t *testing.T) {
 	if result["name"] != "test-instance" {
 		t.Errorf("Expected name 'test-instance', got %v", result["name"])
 	}
-	if result["running"] != false {
-		t.Errorf("Expected running false, got %v", result["running"])
+	if result["status"] != "stopped" {
+		t.Errorf("Expected status 'stopped', got %v", result["status"])
 	}
 
 	// Check that options are included
@@ -210,20 +307,41 @@ func TestMarshalJSON(t *testing.T) {
 	if !ok {
 		t.Error("Expected options to be a map")
 	}
-	if options_map["model"] != "/path/to/model.gguf" {
-		t.Errorf("Expected model '/path/to/model.gguf', got %v", options_map["model"])
+
+	// Check backend type
+	if options_map["backend_type"] != string(backends.BackendTypeLlamaCpp) {
+		t.Errorf("Expected backend_type '%s', got %v", backends.BackendTypeLlamaCpp, options_map["backend_type"])
+	}
+
+	// Check backend options
+	backend_options_data, ok := options_map["backend_options"]
+	if !ok {
+		t.Error("Expected backend_options to be included in JSON")
+	}
+	backend_options_map, ok := backend_options_data.(map[string]any)
+	if !ok {
+		t.Error("Expected backend_options to be a map")
+	}
+	if backend_options_map["model"] != "/path/to/model.gguf" {
+		t.Errorf("Expected model '/path/to/model.gguf', got %v", backend_options_map["model"])
+	}
+	if backend_options_map["port"] != float64(8080) {
+		t.Errorf("Expected port 8080, got %v", backend_options_map["port"])
 	}
 }
 
 func TestUnmarshalJSON(t *testing.T) {
 	jsonData := `{
 		"name": "test-instance",
-		"running": true,
+		"status": "running",
 		"options": {
-			"model": "/path/to/model.gguf",
-			"port": 8080,
 			"auto_restart": false,
-			"max_restarts": 5
+			"max_restarts": 5,
+			"backend_type": "llama_cpp",
+			"backend_options": {
+				"model": "/path/to/model.gguf",
+				"port": 8080
+			}
 		}
 	}`
 
@@ -236,19 +354,25 @@ func TestUnmarshalJSON(t *testing.T) {
 	if inst.Name != "test-instance" {
 		t.Errorf("Expected name 'test-instance', got %q", inst.Name)
 	}
-	if !inst.Running {
-		t.Error("Expected running to be true")
+	if !inst.IsRunning() {
+		t.Error("Expected status to be running")
 	}
 
 	opts := inst.GetOptions()
 	if opts == nil {
 		t.Fatal("Expected options to be set")
 	}
-	if opts.Model != "/path/to/model.gguf" {
-		t.Errorf("Expected model '/path/to/model.gguf', got %q", opts.Model)
+	if opts.BackendType != backends.BackendTypeLlamaCpp {
+		t.Errorf("Expected backend_type '%s', got %s", backends.BackendTypeLlamaCpp, opts.BackendType)
 	}
-	if opts.Port != 8080 {
-		t.Errorf("Expected port 8080, got %d", opts.Port)
+	if opts.LlamaServerOptions == nil {
+		t.Fatal("Expected LlamaServerOptions to be set")
+	}
+	if opts.LlamaServerOptions.Model != "/path/to/model.gguf" {
+		t.Errorf("Expected model '/path/to/model.gguf', got %q", opts.LlamaServerOptions.Model)
+	}
+	if inst.GetPort() != 8080 {
+		t.Errorf("Expected port 8080, got %d", inst.GetPort())
 	}
 	if opts.AutoRestart == nil || *opts.AutoRestart {
 		t.Error("Expected AutoRestart to be false")
@@ -289,6 +413,21 @@ func TestCreateInstanceOptionsValidation(t *testing.T) {
 		},
 	}
 
+	backendConfig := &config.BackendConfig{
+		LlamaCpp: config.BackendSettings{
+			Command: "llama-server",
+			Args:    []string{},
+		},
+		MLX: config.BackendSettings{
+			Command: "mlx_lm.server",
+			Args:    []string{},
+		},
+		VLLM: config.BackendSettings{
+			Command: "vllm",
+			Args:    []string{"serve"},
+		},
+	}
+
 	globalSettings := &config.InstancesConfig{
 		LogsDir: "/tmp/test",
 	}
@@ -298,12 +437,16 @@ func TestCreateInstanceOptionsValidation(t *testing.T) {
 			options := &instance.CreateInstanceOptions{
 				MaxRestarts:  tt.maxRestarts,
 				RestartDelay: tt.restartDelay,
-				LlamaServerOptions: llamacpp.LlamaServerOptions{
+				BackendType:  backends.BackendTypeLlamaCpp,
+				LlamaServerOptions: &llamacpp.LlamaServerOptions{
 					Model: "/path/to/model.gguf",
 				},
 			}
 
-			instance := instance.NewInstance("test", globalSettings, options)
+			// Mock onStatusChange function
+			mockOnStatusChange := func(oldStatus, newStatus instance.InstanceStatus) {}
+
+			instance := instance.NewInstance("test", backendConfig, globalSettings, options, mockOnStatusChange)
 			opts := instance.GetOptions()
 
 			if opts.MaxRestarts == nil {

pkg/instance/lifecycle.go

@@ -5,10 +5,14 @@ import (
 	"fmt"
 	"log"
 	"net/http"
+	"os"
 	"os/exec"
 	"runtime"
 	"syscall"
 	"time"
+
+	"llamactl/pkg/backends"
+	"llamactl/pkg/config"
 )
 
 // Start starts the llama server instance and returns an error if it fails.
@@ -16,7 +20,7 @@ func (i *Process) Start() error {
 	i.mu.Lock()
 	defer i.mu.Unlock()
 
-	if i.Running {
+	if i.IsRunning() {
 		return fmt.Errorf("instance %s is already running", i.Name)
 	}
 
@@ -34,15 +38,20 @@ func (i *Process) Start() error {
 	// Initialize last request time to current time when starting
 	i.lastRequestTime.Store(i.timeProvider.Now().Unix())
 
+	// Create context before building command (needed for CommandContext)
+	i.ctx, i.cancel = context.WithCancel(context.Background())
+
 	// Create log files
 	if err := i.logger.Create(); err != nil {
 		return fmt.Errorf("failed to create log files: %w", err)
 	}
 
-	args := i.options.BuildCommandArgs()
-
-	i.ctx, i.cancel = context.WithCancel(context.Background())
-	i.cmd = exec.CommandContext(i.ctx, "llama-server", args...)
+	// Build command using backend-specific methods
+	cmd, cmdErr := i.buildCommand()
+	if cmdErr != nil {
+		return fmt.Errorf("failed to build command: %w", cmdErr)
+	}
+	i.cmd = cmd
 
 	if runtime.GOOS != "windows" {
 		setProcAttrs(i.cmd)
@@ -65,7 +74,7 @@ func (i *Process) Start() error {
 		return fmt.Errorf("failed to start instance %s: %w", i.Name, err)
 	}
 
-	i.Running = true
+	i.SetStatus(Running)
 
 	// Create channel for monitor completion signaling
 	i.monitorDone = make(chan struct{})
@@ -82,7 +91,7 @@ func (i *Process) Start() error {
 func (i *Process) Stop() error {
 	i.mu.Lock()
 
-	if !i.Running {
+	if !i.IsRunning() {
 		// Even if not running, cancel any pending restart
 		if i.restartCancel != nil {
 			i.restartCancel()
@@ -99,8 +108,8 @@ func (i *Process) Stop() error {
 		i.restartCancel = nil
 	}
 
-	// Set running to false first to signal intentional stop
-	i.Running = false
+	// Set status to stopped first to signal intentional stop
+	i.SetStatus(Stopped)
 
 	// Clean up the proxy
 	i.proxy = nil
@@ -110,19 +119,25 @@ func (i *Process) Stop() error {
 
 	i.mu.Unlock()
 
-	// Stop the process with SIGINT
-	if i.cmd.Process != nil {
+	// Stop the process with SIGINT if cmd exists
+	if i.cmd != nil && i.cmd.Process != nil {
 		if err := i.cmd.Process.Signal(syscall.SIGINT); err != nil {
 			log.Printf("Failed to send SIGINT to instance %s: %v", i.Name, err)
 		}
 	}
 
+	// If no process exists, we can return immediately
+	if i.cmd == nil || monitorDone == nil {
+		i.logger.Close()
+		return nil
+	}
+
 	select {
 	case <-monitorDone:
 		// Process exited normally
 	case <-time.After(30 * time.Second):
 		// Force kill if it doesn't exit within 30 seconds
-		if i.cmd.Process != nil {
+		if i.cmd != nil && i.cmd.Process != nil {
 			killErr := i.cmd.Process.Kill()
 			if killErr != nil {
 				log.Printf("Failed to force kill instance %s: %v", i.Name, killErr)
@@ -144,8 +159,12 @@ func (i *Process) Stop() error {
 	return nil
 }
 
+func (i *Process) LastRequestTime() int64 {
+	return i.lastRequestTime.Load()
+}
+
 func (i *Process) WaitForHealthy(timeout int) error {
-	if !i.Running {
+	if !i.IsRunning() {
 		return fmt.Errorf("instance %s is not running", i.Name)
 	}
 
@@ -163,11 +182,29 @@ func (i *Process) WaitForHealthy(timeout int) error {
 	}
 
 	// Build the health check URL directly
-	host := opts.Host
+	var host string
+	var port int
+	switch opts.BackendType {
+	case backends.BackendTypeLlamaCpp:
+		if opts.LlamaServerOptions != nil {
+			host = opts.LlamaServerOptions.Host
+			port = opts.LlamaServerOptions.Port
+		}
+	case backends.BackendTypeMlxLm:
+		if opts.MlxServerOptions != nil {
+			host = opts.MlxServerOptions.Host
+			port = opts.MlxServerOptions.Port
+		}
+	case backends.BackendTypeVllm:
+		if opts.VllmServerOptions != nil {
+			host = opts.VllmServerOptions.Host
+			port = opts.VllmServerOptions.Port
+		}
+	}
 	if host == "" {
 		host = "localhost"
 	}
-	healthURL := fmt.Sprintf("http://%s:%d/health", host, opts.Port)
+	healthURL := fmt.Sprintf("http://%s:%d/health", host, port)
 
 	// Create a dedicated HTTP client for health checks
 	client := &http.Client{
@@ -227,12 +264,12 @@ func (i *Process) monitorProcess() {
 	i.mu.Lock()
 
 	// Check if the instance was intentionally stopped
-	if !i.Running {
+	if !i.IsRunning() {
 		i.mu.Unlock()
 		return
 	}
 
-	i.Running = false
+	i.SetStatus(Stopped)
 	i.logger.Close()
 
 	// Cancel any existing restart context since we're handling a new exit
@@ -257,6 +294,7 @@ func (i *Process) handleRestart() {
 	// Validate restart conditions and get safe parameters
 	shouldRestart, maxRestarts, restartDelay := i.validateRestartConditions()
 	if !shouldRestart {
+		i.SetStatus(Failed)
 		i.mu.Unlock()
 		return
 	}
@@ -327,3 +365,53 @@ func (i *Process) validateRestartConditions() (shouldRestart bool, maxRestarts i
 
 	return true, maxRestarts, restartDelay
 }
+
+// buildCommand builds the command to execute using backend-specific logic
+func (i *Process) buildCommand() (*exec.Cmd, error) {
+	// Get backend configuration
+	backendConfig, err := i.getBackendConfig()
+	if err != nil {
+		return nil, err
+	}
+
+	// Build the environment variables
+	env := i.options.BuildEnvironment(backendConfig)
+
+	// Get the command to execute
+	command := i.options.GetCommand(backendConfig)
+
+	// Build command arguments
+	args := i.options.BuildCommandArgs(backendConfig)
+
+	// Create the exec.Cmd
+	cmd := exec.CommandContext(i.ctx, command, args...)
+
+	// Start with host environment variables
+	cmd.Env = os.Environ()
+
+	// Add/override with backend-specific environment variables
+	for k, v := range env {
+		cmd.Env = append(cmd.Env, fmt.Sprintf("%s=%s", k, v))
+	}
+
+	return cmd, nil
+}
+
+// getBackendConfig resolves the backend configuration for the current instance
+func (i *Process) getBackendConfig() (*config.BackendSettings, error) {
+	var backendTypeStr string
+
+	switch i.options.BackendType {
+	case backends.BackendTypeLlamaCpp:
+		backendTypeStr = "llama-cpp"
+	case backends.BackendTypeMlxLm:
+		backendTypeStr = "mlx"
+	case backends.BackendTypeVllm:
+		backendTypeStr = "vllm"
+	default:
+		return nil, fmt.Errorf("unsupported backend type: %s", i.options.BackendType)
+	}
+
+	settings := i.globalBackendSettings.GetBackendSettings(backendTypeStr)
+	return &settings, nil
+}

pkg/instance/options.go

@@ -0,0 +1,265 @@
+package instance
+
+import (
+	"encoding/json"
+	"fmt"
+	"llamactl/pkg/backends"
+	"llamactl/pkg/backends/llamacpp"
+	"llamactl/pkg/backends/mlx"
+	"llamactl/pkg/backends/vllm"
+	"llamactl/pkg/config"
+	"log"
+	"maps"
+)
+
+type CreateInstanceOptions struct {
+	// Auto restart
+	AutoRestart  *bool `json:"auto_restart,omitempty"`
+	MaxRestarts  *int  `json:"max_restarts,omitempty"`
+	RestartDelay *int  `json:"restart_delay,omitempty"` // seconds
+	// On demand start
+	OnDemandStart *bool `json:"on_demand_start,omitempty"`
+	// Idle timeout
+	IdleTimeout *int `json:"idle_timeout,omitempty"` // minutes
+	//Environment variables
+	Environment map[string]string `json:"environment,omitempty"`
+
+	BackendType    backends.BackendType `json:"backend_type"`
+	BackendOptions map[string]any       `json:"backend_options,omitempty"`
+
+	// Backend-specific options
+	LlamaServerOptions *llamacpp.LlamaServerOptions `json:"-"`
+	MlxServerOptions   *mlx.MlxServerOptions        `json:"-"`
+	VllmServerOptions  *vllm.VllmServerOptions      `json:"-"`
+}
+
+// UnmarshalJSON implements custom JSON unmarshaling for CreateInstanceOptions
+func (c *CreateInstanceOptions) UnmarshalJSON(data []byte) error {
+	// Use anonymous struct to avoid recursion
+	type Alias CreateInstanceOptions
+	aux := &struct {
+		*Alias
+	}{
+		Alias: (*Alias)(c),
+	}
+
+	if err := json.Unmarshal(data, aux); err != nil {
+		return err
+	}
+
+	// Parse backend-specific options
+	switch c.BackendType {
+	case backends.BackendTypeLlamaCpp:
+		if c.BackendOptions != nil {
+			// Convert map to JSON and then unmarshal to LlamaServerOptions
+			optionsData, err := json.Marshal(c.BackendOptions)
+			if err != nil {
+				return fmt.Errorf("failed to marshal backend options: %w", err)
+			}
+
+			c.LlamaServerOptions = &llamacpp.LlamaServerOptions{}
+			if err := json.Unmarshal(optionsData, c.LlamaServerOptions); err != nil {
+				return fmt.Errorf("failed to unmarshal llama.cpp options: %w", err)
+			}
+		}
+	case backends.BackendTypeMlxLm:
+		if c.BackendOptions != nil {
+			optionsData, err := json.Marshal(c.BackendOptions)
+			if err != nil {
+				return fmt.Errorf("failed to marshal backend options: %w", err)
+			}
+
+			c.MlxServerOptions = &mlx.MlxServerOptions{}
+			if err := json.Unmarshal(optionsData, c.MlxServerOptions); err != nil {
+				return fmt.Errorf("failed to unmarshal MLX options: %w", err)
+			}
+		}
+	case backends.BackendTypeVllm:
+		if c.BackendOptions != nil {
+			optionsData, err := json.Marshal(c.BackendOptions)
+			if err != nil {
+				return fmt.Errorf("failed to marshal backend options: %w", err)
+			}
+
+			c.VllmServerOptions = &vllm.VllmServerOptions{}
+			if err := json.Unmarshal(optionsData, c.VllmServerOptions); err != nil {
+				return fmt.Errorf("failed to unmarshal vLLM options: %w", err)
+			}
+		}
+	default:
+		return fmt.Errorf("unknown backend type: %s", c.BackendType)
+	}
+
+	return nil
+}
+
+// MarshalJSON implements custom JSON marshaling for CreateInstanceOptions
+func (c *CreateInstanceOptions) MarshalJSON() ([]byte, error) {
+	// Use anonymous struct to avoid recursion
+	type Alias CreateInstanceOptions
+	aux := struct {
+		*Alias
+	}{
+		Alias: (*Alias)(c),
+	}
+
+	// Convert backend-specific options back to BackendOptions map for JSON
+	switch c.BackendType {
+	case backends.BackendTypeLlamaCpp:
+		if c.LlamaServerOptions != nil {
+			data, err := json.Marshal(c.LlamaServerOptions)
+			if err != nil {
+				return nil, fmt.Errorf("failed to marshal llama server options: %w", err)
+			}
+
+			var backendOpts map[string]any
+			if err := json.Unmarshal(data, &backendOpts); err != nil {
+				return nil, fmt.Errorf("failed to unmarshal to map: %w", err)
+			}
+
+			aux.BackendOptions = backendOpts
+		}
+	case backends.BackendTypeMlxLm:
+		if c.MlxServerOptions != nil {
+			data, err := json.Marshal(c.MlxServerOptions)
+			if err != nil {
+				return nil, fmt.Errorf("failed to marshal MLX server options: %w", err)
+			}
+
+			var backendOpts map[string]any
+			if err := json.Unmarshal(data, &backendOpts); err != nil {
+				return nil, fmt.Errorf("failed to unmarshal to map: %w", err)
+			}
+
+			aux.BackendOptions = backendOpts
+		}
+	case backends.BackendTypeVllm:
+		if c.VllmServerOptions != nil {
+			data, err := json.Marshal(c.VllmServerOptions)
+			if err != nil {
+				return nil, fmt.Errorf("failed to marshal vLLM server options: %w", err)
+			}
+
+			var backendOpts map[string]any
+			if err := json.Unmarshal(data, &backendOpts); err != nil {
+				return nil, fmt.Errorf("failed to unmarshal to map: %w", err)
+			}
+
+			aux.BackendOptions = backendOpts
+		}
+	}
+
+	return json.Marshal(aux)
+}
+
+// ValidateAndApplyDefaults validates the instance options and applies constraints
+func (c *CreateInstanceOptions) ValidateAndApplyDefaults(name string, globalSettings *config.InstancesConfig) {
+	// Validate and apply constraints
+	if c.MaxRestarts != nil && *c.MaxRestarts < 0 {
+		log.Printf("Instance %s MaxRestarts value (%d) cannot be negative, setting to 0", name, *c.MaxRestarts)
+		*c.MaxRestarts = 0
+	}
+
+	if c.RestartDelay != nil && *c.RestartDelay < 0 {
+		log.Printf("Instance %s RestartDelay value (%d) cannot be negative, setting to 0 seconds", name, *c.RestartDelay)
+		*c.RestartDelay = 0
+	}
+
+	if c.IdleTimeout != nil && *c.IdleTimeout < 0 {
+		log.Printf("Instance %s IdleTimeout value (%d) cannot be negative, setting to 0 minutes", name, *c.IdleTimeout)
+		*c.IdleTimeout = 0
+	}
+
+	// Apply defaults from global settings for nil fields
+	if globalSettings != nil {
+		if c.AutoRestart == nil {
+			c.AutoRestart = &globalSettings.DefaultAutoRestart
+		}
+		if c.MaxRestarts == nil {
+			c.MaxRestarts = &globalSettings.DefaultMaxRestarts
+		}
+		if c.RestartDelay == nil {
+			c.RestartDelay = &globalSettings.DefaultRestartDelay
+		}
+		if c.OnDemandStart == nil {
+			c.OnDemandStart = &globalSettings.DefaultOnDemandStart
+		}
+		if c.IdleTimeout == nil {
+			defaultIdleTimeout := 0
+			c.IdleTimeout = &defaultIdleTimeout
+		}
+	}
+}
+
+func (c *CreateInstanceOptions) GetCommand(backendConfig *config.BackendSettings) string {
+
+	if backendConfig.Docker != nil && backendConfig.Docker.Enabled && c.BackendType != backends.BackendTypeMlxLm {
+		return "docker"
+	}
+
+	return backendConfig.Command
+}
+
+// BuildCommandArgs builds command line arguments for the backend
+func (c *CreateInstanceOptions) BuildCommandArgs(backendConfig *config.BackendSettings) []string {
+
+	var args []string
+
+	if backendConfig.Docker != nil && backendConfig.Docker.Enabled && c.BackendType != backends.BackendTypeMlxLm {
+		// For Docker, start with Docker args
+		args = append(args, backendConfig.Docker.Args...)
+		args = append(args, backendConfig.Docker.Image)
+
+		switch c.BackendType {
+		case backends.BackendTypeLlamaCpp:
+			if c.LlamaServerOptions != nil {
+				args = append(args, c.LlamaServerOptions.BuildDockerArgs()...)
+			}
+		case backends.BackendTypeVllm:
+			if c.VllmServerOptions != nil {
+				args = append(args, c.VllmServerOptions.BuildDockerArgs()...)
+			}
+		}
+
+	} else {
+		// For native execution, start with backend args
+		args = append(args, backendConfig.Args...)
+
+		switch c.BackendType {
+		case backends.BackendTypeLlamaCpp:
+			if c.LlamaServerOptions != nil {
+				args = append(args, c.LlamaServerOptions.BuildCommandArgs()...)
+			}
+		case backends.BackendTypeMlxLm:
+			if c.MlxServerOptions != nil {
+				args = append(args, c.MlxServerOptions.BuildCommandArgs()...)
+			}
+		case backends.BackendTypeVllm:
+			if c.VllmServerOptions != nil {
+				args = append(args, c.VllmServerOptions.BuildCommandArgs()...)
+			}
+		}
+	}
+
+	return args
+}
+
+func (c *CreateInstanceOptions) BuildEnvironment(backendConfig *config.BackendSettings) map[string]string {
+	env := map[string]string{}
+
+	if backendConfig.Environment != nil {
+		maps.Copy(env, backendConfig.Environment)
+	}
+
+	if backendConfig.Docker != nil && backendConfig.Docker.Enabled && c.BackendType != backends.BackendTypeMlxLm {
+		if backendConfig.Docker.Environment != nil {
+			maps.Copy(env, backendConfig.Docker.Environment)
+		}
+	}
+
+	if c.Environment != nil {
+		maps.Copy(env, c.Environment)
+	}
+
+	return env
+}

pkg/instance/status.go

@@ -0,0 +1,70 @@
+package instance
+
+import (
+	"encoding/json"
+	"log"
+)
+
+// Enum for instance status
+type InstanceStatus int
+
+const (
+	Stopped InstanceStatus = iota
+	Running
+	Failed
+)
+
+var nameToStatus = map[string]InstanceStatus{
+	"stopped": Stopped,
+	"running": Running,
+	"failed":  Failed,
+}
+
+var statusToName = map[InstanceStatus]string{
+	Stopped: "stopped",
+	Running: "running",
+	Failed:  "failed",
+}
+
+func (p *Process) SetStatus(status InstanceStatus) {
+	oldStatus := p.Status
+	p.Status = status
+
+	if p.onStatusChange != nil {
+		p.onStatusChange(oldStatus, status)
+	}
+}
+
+func (p *Process) GetStatus() InstanceStatus {
+	return p.Status
+}
+
+// IsRunning returns true if the status is Running
+func (p *Process) IsRunning() bool {
+	return p.Status == Running
+}
+
+func (s InstanceStatus) MarshalJSON() ([]byte, error) {
+	name, ok := statusToName[s]
+	if !ok {
+		name = "stopped" // Default to "stopped" for unknown status
+	}
+	return json.Marshal(name)
+}
+
+// UnmarshalJSON implements json.Unmarshaler
+func (s *InstanceStatus) UnmarshalJSON(data []byte) error {
+	var str string
+	if err := json.Unmarshal(data, &str); err != nil {
+		return err
+	}
+
+	status, ok := nameToStatus[str]
+	if !ok {
+		log.Printf("Unknown instance status: %s", str)
+		status = Stopped // Default to Stopped on unknown status
+	}
+
+	*s = status
+	return nil
+}

pkg/instance/timeout.go

@@ -13,7 +13,7 @@ func (i *Process) ShouldTimeout() bool {
 	i.mu.RLock()
 	defer i.mu.RUnlock()
 
-	if !i.Running || i.options.IdleTimeout == nil || *i.options.IdleTimeout <= 0 {
+	if !i.IsRunning() || i.options.IdleTimeout == nil || *i.options.IdleTimeout <= 0 {
 		return false
 	}
 

pkg/instance/timeout_test.go

@@ -1,6 +1,7 @@
 package instance_test
 
 import (
+	"llamactl/pkg/backends"
 	"llamactl/pkg/backends/llamacpp"
 	"llamactl/pkg/config"
 	"llamactl/pkg/instance"
@@ -32,23 +33,45 @@ func (m *MockTimeProvider) SetTime(t time.Time) {
 // Timeout-related tests
 
 func TestUpdateLastRequestTime(t *testing.T) {
+	backendConfig := &config.BackendConfig{
+		LlamaCpp: config.BackendSettings{
+			Command: "llama-server",
+		},
+		MLX: config.BackendSettings{
+			Command: "mlx_lm.server",
+		},
+	}
+
 	globalSettings := &config.InstancesConfig{
 		LogsDir: "/tmp/test",
 	}
 
 	options := &instance.CreateInstanceOptions{
-		LlamaServerOptions: llamacpp.LlamaServerOptions{
+		BackendType: backends.BackendTypeLlamaCpp,
+		LlamaServerOptions: &llamacpp.LlamaServerOptions{
 			Model: "/path/to/model.gguf",
 		},
 	}
 
-	inst := instance.NewInstance("test-instance", globalSettings, options)
+	// Mock onStatusChange function
+	mockOnStatusChange := func(oldStatus, newStatus instance.InstanceStatus) {}
+
+	inst := instance.NewInstance("test-instance", backendConfig, globalSettings, options, mockOnStatusChange)
 
 	// Test that UpdateLastRequestTime doesn't panic
 	inst.UpdateLastRequestTime()
 }
 
 func TestShouldTimeout_NotRunning(t *testing.T) {
+	backendConfig := &config.BackendConfig{
+		LlamaCpp: config.BackendSettings{
+			Command: "llama-server",
+		},
+		MLX: config.BackendSettings{
+			Command: "mlx_lm.server",
+		},
+	}
+
 	globalSettings := &config.InstancesConfig{
 		LogsDir: "/tmp/test",
 	}
@@ -56,12 +79,16 @@ func TestShouldTimeout_NotRunning(t *testing.T) {
 	idleTimeout := 1 // 1 minute
 	options := &instance.CreateInstanceOptions{
 		IdleTimeout: &idleTimeout,
-		LlamaServerOptions: llamacpp.LlamaServerOptions{
+		BackendType: backends.BackendTypeLlamaCpp,
+		LlamaServerOptions: &llamacpp.LlamaServerOptions{
 			Model: "/path/to/model.gguf",
 		},
 	}
 
-	inst := instance.NewInstance("test-instance", globalSettings, options)
+	// Mock onStatusChange function
+	mockOnStatusChange := func(oldStatus, newStatus instance.InstanceStatus) {}
+
+	inst := instance.NewInstance("test-instance", backendConfig, globalSettings, options, mockOnStatusChange)
 
 	// Instance is not running, should not timeout regardless of configuration
 	if inst.ShouldTimeout() {
@@ -70,6 +97,15 @@ func TestShouldTimeout_NotRunning(t *testing.T) {
 }
 
 func TestShouldTimeout_NoTimeoutConfigured(t *testing.T) {
+	backendConfig := &config.BackendConfig{
+		LlamaCpp: config.BackendSettings{
+			Command: "llama-server",
+		},
+		MLX: config.BackendSettings{
+			Command: "mlx_lm.server",
+		},
+	}
+
 	globalSettings := &config.InstancesConfig{
 		LogsDir: "/tmp/test",
 	}
@@ -85,16 +121,20 @@ func TestShouldTimeout_NoTimeoutConfigured(t *testing.T) {
 
 	for _, tt := range tests {
 		t.Run(tt.name, func(t *testing.T) {
+			// Mock onStatusChange function
+			mockOnStatusChange := func(oldStatus, newStatus instance.InstanceStatus) {}
+
 			options := &instance.CreateInstanceOptions{
 				IdleTimeout: tt.idleTimeout,
-				LlamaServerOptions: llamacpp.LlamaServerOptions{
+				BackendType: backends.BackendTypeLlamaCpp,
+				LlamaServerOptions: &llamacpp.LlamaServerOptions{
 					Model: "/path/to/model.gguf",
 				},
 			}
 
-			inst := instance.NewInstance("test-instance", globalSettings, options)
+			inst := instance.NewInstance("test-instance", backendConfig, globalSettings, options, mockOnStatusChange)
 			// Simulate running state
-			inst.Running = true
+			inst.SetStatus(instance.Running)
 
 			if inst.ShouldTimeout() {
 				t.Errorf("Instance with %s should not timeout", tt.name)
@@ -104,6 +144,15 @@ func TestShouldTimeout_NoTimeoutConfigured(t *testing.T) {
 }
 
 func TestShouldTimeout_WithinTimeLimit(t *testing.T) {
+	backendConfig := &config.BackendConfig{
+		LlamaCpp: config.BackendSettings{
+			Command: "llama-server",
+		},
+		MLX: config.BackendSettings{
+			Command: "mlx_lm.server",
+		},
+	}
+
 	globalSettings := &config.InstancesConfig{
 		LogsDir: "/tmp/test",
 	}
@@ -111,13 +160,17 @@ func TestShouldTimeout_WithinTimeLimit(t *testing.T) {
 	idleTimeout := 5 // 5 minutes
 	options := &instance.CreateInstanceOptions{
 		IdleTimeout: &idleTimeout,
-		LlamaServerOptions: llamacpp.LlamaServerOptions{
+		BackendType: backends.BackendTypeLlamaCpp,
+		LlamaServerOptions: &llamacpp.LlamaServerOptions{
 			Model: "/path/to/model.gguf",
 		},
 	}
 
-	inst := instance.NewInstance("test-instance", globalSettings, options)
-	inst.Running = true
+	// Mock onStatusChange function
+	mockOnStatusChange := func(oldStatus, newStatus instance.InstanceStatus) {}
+
+	inst := instance.NewInstance("test-instance", backendConfig, globalSettings, options, mockOnStatusChange)
+	inst.SetStatus(instance.Running)
 
 	// Update last request time to now
 	inst.UpdateLastRequestTime()
@@ -129,6 +182,15 @@ func TestShouldTimeout_WithinTimeLimit(t *testing.T) {
 }
 
 func TestShouldTimeout_ExceedsTimeLimit(t *testing.T) {
+	backendConfig := &config.BackendConfig{
+		LlamaCpp: config.BackendSettings{
+			Command: "llama-server",
+		},
+		MLX: config.BackendSettings{
+			Command: "mlx_lm.server",
+		},
+	}
+
 	globalSettings := &config.InstancesConfig{
 		LogsDir: "/tmp/test",
 	}
@@ -136,13 +198,17 @@ func TestShouldTimeout_ExceedsTimeLimit(t *testing.T) {
 	idleTimeout := 1 // 1 minute
 	options := &instance.CreateInstanceOptions{
 		IdleTimeout: &idleTimeout,
-		LlamaServerOptions: llamacpp.LlamaServerOptions{
+		BackendType: backends.BackendTypeLlamaCpp,
+		LlamaServerOptions: &llamacpp.LlamaServerOptions{
 			Model: "/path/to/model.gguf",
 		},
 	}
 
-	inst := instance.NewInstance("test-instance", globalSettings, options)
-	inst.Running = true
+	// Mock onStatusChange function
+	mockOnStatusChange := func(oldStatus, newStatus instance.InstanceStatus) {}
+
+	inst := instance.NewInstance("test-instance", backendConfig, globalSettings, options, mockOnStatusChange)
+	inst.SetStatus(instance.Running)
 
 	// Use MockTimeProvider to simulate old last request time
 	mockTime := NewMockTimeProvider(time.Now())
@@ -160,6 +226,15 @@ func TestShouldTimeout_ExceedsTimeLimit(t *testing.T) {
 }
 
 func TestTimeoutConfiguration_Validation(t *testing.T) {
+	backendConfig := &config.BackendConfig{
+		LlamaCpp: config.BackendSettings{
+			Command: "llama-server",
+		},
+		MLX: config.BackendSettings{
+			Command: "mlx_lm.server",
+		},
+	}
+
 	globalSettings := &config.InstancesConfig{
 		LogsDir: "/tmp/test",
 	}
@@ -179,12 +254,16 @@ func TestTimeoutConfiguration_Validation(t *testing.T) {
 		t.Run(tt.name, func(t *testing.T) {
 			options := &instance.CreateInstanceOptions{
 				IdleTimeout: tt.inputTimeout,
-				LlamaServerOptions: llamacpp.LlamaServerOptions{
+				BackendType: backends.BackendTypeLlamaCpp,
+				LlamaServerOptions: &llamacpp.LlamaServerOptions{
 					Model: "/path/to/model.gguf",
 				},
 			}
 
-			inst := instance.NewInstance("test-instance", globalSettings, options)
+			// Mock onStatusChange function
+			mockOnStatusChange := func(oldStatus, newStatus instance.InstanceStatus) {}
+
+			inst := instance.NewInstance("test-instance", backendConfig, globalSettings, options, mockOnStatusChange)
 			opts := inst.GetOptions()
 
 			if opts.IdleTimeout == nil || *opts.IdleTimeout != tt.expectedTimeout {

pkg/manager/manager.go

@@ -21,17 +21,21 @@ type InstanceManager interface {
 	UpdateInstance(name string, options *instance.CreateInstanceOptions) (*instance.Process, error)
 	DeleteInstance(name string) error
 	StartInstance(name string) (*instance.Process, error)
+	IsMaxRunningInstancesReached() bool
 	StopInstance(name string) (*instance.Process, error)
+	EvictLRUInstance() error
 	RestartInstance(name string) (*instance.Process, error)
 	GetInstanceLogs(name string) (string, error)
 	Shutdown()
 }
 
 type instanceManager struct {
-	mu              sync.RWMutex
-	instances       map[string]*instance.Process
-	ports           map[int]bool
-	instancesConfig config.InstancesConfig
+	mu               sync.RWMutex
+	instances        map[string]*instance.Process
+	runningInstances map[string]struct{}
+	ports            map[int]bool
+	instancesConfig  config.InstancesConfig
+	backendsConfig   config.BackendConfig
 
 	// Timeout checker
 	timeoutChecker *time.Ticker
@@ -41,14 +45,16 @@ type instanceManager struct {
 }
 
 // NewInstanceManager creates a new instance of InstanceManager.
-func NewInstanceManager(instancesConfig config.InstancesConfig) InstanceManager {
+func NewInstanceManager(backendsConfig config.BackendConfig, instancesConfig config.InstancesConfig) InstanceManager {
 	if instancesConfig.TimeoutCheckInterval <= 0 {
 		instancesConfig.TimeoutCheckInterval = 5 // Default to 5 minutes if not set
 	}
 	im := &instanceManager{
-		instances:       make(map[string]*instance.Process),
-		ports:           make(map[int]bool),
-		instancesConfig: instancesConfig,
+		instances:        make(map[string]*instance.Process),
+		runningInstances: make(map[string]struct{}),
+		ports:            make(map[int]bool),
+		instancesConfig:  instancesConfig,
+		backendsConfig:   backendsConfig,
 
 		timeoutChecker: time.NewTicker(time.Duration(instancesConfig.TimeoutCheckInterval) * time.Minute),
 		shutdownChan:   make(chan struct{}),
@@ -121,10 +127,10 @@ func (im *instanceManager) persistInstance(instance *instance.Process) error {
 
 func (im *instanceManager) Shutdown() {
 	im.mu.Lock()
-	defer im.mu.Unlock()
 
 	// Check if already shutdown
 	if im.isShutdown {
+		im.mu.Unlock()
 		return
 	}
 	im.isShutdown = true
@@ -132,27 +138,32 @@ func (im *instanceManager) Shutdown() {
 	// Signal the timeout checker to stop
 	close(im.shutdownChan)
 
-	// Release lock temporarily to wait for goroutine
+	// Create a list of running instances to stop
+	var runningInstances []*instance.Process
+	var runningNames []string
+	for name, inst := range im.instances {
+		if inst.IsRunning() {
+			runningInstances = append(runningInstances, inst)
+			runningNames = append(runningNames, name)
+		}
+	}
+
+	// Release lock before stopping instances to avoid deadlock
 	im.mu.Unlock()
+
 	// Wait for the timeout checker goroutine to actually stop
 	<-im.shutdownDone
-	// Reacquire lock
-	im.mu.Lock()
 
 	// Now stop the ticker
 	if im.timeoutChecker != nil {
 		im.timeoutChecker.Stop()
 	}
 
+	// Stop instances without holding the manager lock
 	var wg sync.WaitGroup
-	wg.Add(len(im.instances))
-
-	for name, inst := range im.instances {
-		if !inst.Running {
-			wg.Done() // If instance is not running, just mark it as done
-			continue
-		}
+	wg.Add(len(runningInstances))
 
+	for i, inst := range runningInstances {
 		go func(name string, inst *instance.Process) {
 			defer wg.Done()
 			fmt.Printf("Stopping instance %s...\n", name)
@@ -160,7 +171,7 @@ func (im *instanceManager) Shutdown() {
 			if err := inst.Stop(); err != nil {
 				fmt.Printf("Error stopping instance %s: %v\n", name, err)
 			}
-		}(name, inst)
+		}(runningNames[i], inst)
 	}
 
 	wg.Wait()
@@ -227,16 +238,20 @@ func (im *instanceManager) loadInstance(name, path string) error {
 		return fmt.Errorf("instance name mismatch: file=%s, instance.Name=%s", name, persistedInstance.Name)
 	}
 
+	statusCallback := func(oldStatus, newStatus instance.InstanceStatus) {
+		im.onStatusChange(persistedInstance.Name, oldStatus, newStatus)
+	}
+
 	// Create new inst using NewInstance (handles validation, defaults, setup)
-	inst := instance.NewInstance(name, &im.instancesConfig, persistedInstance.GetOptions())
+	inst := instance.NewInstance(name, &im.backendsConfig, &im.instancesConfig, persistedInstance.GetOptions(), statusCallback)
 
 	// Restore persisted fields that NewInstance doesn't set
 	inst.Created = persistedInstance.Created
-	inst.Running = persistedInstance.Running
+	inst.SetStatus(persistedInstance.Status)
 
 	// Check for port conflicts and add to maps
-	if inst.GetOptions() != nil && inst.GetOptions().Port > 0 {
-		port := inst.GetOptions().Port
+	if inst.GetPort() > 0 {
+		port := inst.GetPort()
 		if im.ports[port] {
 			return fmt.Errorf("port conflict: instance %s wants port %d which is already in use", name, port)
 		}
@@ -248,25 +263,49 @@ func (im *instanceManager) loadInstance(name, path string) error {
 }
 
 // autoStartInstances starts instances that were running when persisted and have auto-restart enabled
+// For instances with auto-restart disabled, it sets their status to Stopped
 func (im *instanceManager) autoStartInstances() {
 	im.mu.RLock()
 	var instancesToStart []*instance.Process
+	var instancesToStop []*instance.Process
 	for _, inst := range im.instances {
-		if inst.Running && // Was running when persisted
+		if inst.IsRunning() && // Was running when persisted
 			inst.GetOptions() != nil &&
-			inst.GetOptions().AutoRestart != nil &&
-			*inst.GetOptions().AutoRestart {
-			instancesToStart = append(instancesToStart, inst)
+			inst.GetOptions().AutoRestart != nil {
+			if *inst.GetOptions().AutoRestart {
+				instancesToStart = append(instancesToStart, inst)
+			} else {
+				// Instance was running but auto-restart is disabled, mark as stopped
+				instancesToStop = append(instancesToStop, inst)
+			}
 		}
 	}
 	im.mu.RUnlock()
 
+	// Stop instances that have auto-restart disabled
+	for _, inst := range instancesToStop {
+		log.Printf("Instance %s was running but auto-restart is disabled, setting status to stopped", inst.Name)
+		inst.SetStatus(instance.Stopped)
+	}
+
+	// Start instances that have auto-restart enabled
 	for _, inst := range instancesToStart {
 		log.Printf("Auto-starting instance %s", inst.Name)
 		// Reset running state before starting (since Start() expects stopped instance)
-		inst.Running = false
+		inst.SetStatus(instance.Stopped)
 		if err := inst.Start(); err != nil {
 			log.Printf("Failed to auto-start instance %s: %v", inst.Name, err)
 		}
 	}
 }
+
+func (im *instanceManager) onStatusChange(name string, oldStatus, newStatus instance.InstanceStatus) {
+	im.mu.Lock()
+	defer im.mu.Unlock()
+
+	if newStatus == instance.Running {
+		im.runningInstances[name] = struct{}{}
+	} else {
+		delete(im.runningInstances, name)
+	}
+}

pkg/manager/manager_test.go

@@ -2,6 +2,7 @@ package manager_test
 
 import (
 	"fmt"
+	"llamactl/pkg/backends"
 	"llamactl/pkg/backends/llamacpp"
 	"llamactl/pkg/config"
 	"llamactl/pkg/instance"
@@ -11,28 +12,35 @@ import (
 	"strings"
 	"sync"
 	"testing"
-	"time"
 )
 
 func TestNewInstanceManager(t *testing.T) {
+	backendConfig := config.BackendConfig{
+		LlamaCpp: config.BackendSettings{
+			Command: "llama-server",
+		},
+		MLX: config.BackendSettings{
+			Command: "mlx_lm.server",
+		},
+	}
+
 	cfg := config.InstancesConfig{
 		PortRange:            [2]int{8000, 9000},
 		LogsDir:              "/tmp/test",
 		MaxInstances:         5,
-		LlamaExecutable:      "llama-server",
 		DefaultAutoRestart:   true,
 		DefaultMaxRestarts:   3,
 		DefaultRestartDelay:  5,
 		TimeoutCheckInterval: 5,
 	}
 
-	manager := manager.NewInstanceManager(cfg)
-	if manager == nil {
+	mgr := manager.NewInstanceManager(backendConfig, cfg)
+	if mgr == nil {
 		t.Fatal("NewInstanceManager returned nil")
 	}
 
 	// Test initial state
-	instances, err := manager.ListInstances()
+	instances, err := mgr.ListInstances()
 	if err != nil {
 		t.Fatalf("ListInstances failed: %v", err)
 	}
@@ -41,220 +49,18 @@ func TestNewInstanceManager(t *testing.T) {
 	}
 }
 
-func TestCreateInstance_Success(t *testing.T) {
-	manager := createTestManager()
-
-	options := &instance.CreateInstanceOptions{
-		LlamaServerOptions: llamacpp.LlamaServerOptions{
-			Model: "/path/to/model.gguf",
-			Port:  8080,
-		},
-	}
-
-	inst, err := manager.CreateInstance("test-instance", options)
-	if err != nil {
-		t.Fatalf("CreateInstance failed: %v", err)
-	}
-
-	if inst.Name != "test-instance" {
-		t.Errorf("Expected instance name 'test-instance', got %q", inst.Name)
-	}
-	if inst.Running {
-		t.Error("New instance should not be running")
-	}
-	if inst.GetOptions().Port != 8080 {
-		t.Errorf("Expected port 8080, got %d", inst.GetOptions().Port)
-	}
-}
-
-func TestCreateInstance_ValidationAndLimits(t *testing.T) {
-	// Test duplicate names
-	mngr := createTestManager()
-	options := &instance.CreateInstanceOptions{
-		LlamaServerOptions: llamacpp.LlamaServerOptions{
-			Model: "/path/to/model.gguf",
-		},
-	}
-
-	_, err := mngr.CreateInstance("test-instance", options)
-	if err != nil {
-		t.Fatalf("First CreateInstance failed: %v", err)
-	}
-
-	// Try to create duplicate
-	_, err = mngr.CreateInstance("test-instance", options)
-	if err == nil {
-		t.Error("Expected error for duplicate instance name")
-	}
-	if !strings.Contains(err.Error(), "already exists") {
-		t.Errorf("Expected duplicate name error, got: %v", err)
-	}
-
-	// Test max instances limit
-	cfg := config.InstancesConfig{
-		PortRange:            [2]int{8000, 9000},
-		MaxInstances:         1, // Very low limit for testing
-		TimeoutCheckInterval: 5,
-	}
-	limitedManager := manager.NewInstanceManager(cfg)
-
-	_, err = limitedManager.CreateInstance("instance1", options)
-	if err != nil {
-		t.Fatalf("CreateInstance 1 failed: %v", err)
-	}
-
-	// This should fail due to max instances limit
-	_, err = limitedManager.CreateInstance("instance2", options)
-	if err == nil {
-		t.Error("Expected error when exceeding max instances limit")
-	}
-	if !strings.Contains(err.Error(), "maximum number of instances") {
-		t.Errorf("Expected max instances error, got: %v", err)
-	}
-}
-
-func TestPortManagement(t *testing.T) {
-	manager := createTestManager()
-
-	// Test auto port assignment
-	options1 := &instance.CreateInstanceOptions{
-		LlamaServerOptions: llamacpp.LlamaServerOptions{
-			Model: "/path/to/model.gguf",
-		},
-	}
-
-	inst1, err := manager.CreateInstance("instance1", options1)
-	if err != nil {
-		t.Fatalf("CreateInstance failed: %v", err)
-	}
-
-	port1 := inst1.GetOptions().Port
-	if port1 < 8000 || port1 > 9000 {
-		t.Errorf("Expected port in range 8000-9000, got %d", port1)
-	}
-
-	// Test port conflict detection
-	options2 := &instance.CreateInstanceOptions{
-		LlamaServerOptions: llamacpp.LlamaServerOptions{
-			Model: "/path/to/model2.gguf",
-			Port:  port1, // Same port - should conflict
-		},
-	}
-
-	_, err = manager.CreateInstance("instance2", options2)
-	if err == nil {
-		t.Error("Expected error for port conflict")
-	}
-	if !strings.Contains(err.Error(), "port") && !strings.Contains(err.Error(), "in use") {
-		t.Errorf("Expected port conflict error, got: %v", err)
-	}
-
-	// Test port release on deletion
-	specificPort := 8080
-	options3 := &instance.CreateInstanceOptions{
-		LlamaServerOptions: llamacpp.LlamaServerOptions{
-			Model: "/path/to/model.gguf",
-			Port:  specificPort,
-		},
-	}
-
-	_, err = manager.CreateInstance("port-test", options3)
-	if err != nil {
-		t.Fatalf("CreateInstance failed: %v", err)
-	}
-
-	err = manager.DeleteInstance("port-test")
-	if err != nil {
-		t.Fatalf("DeleteInstance failed: %v", err)
-	}
-
-	// Should be able to create new instance with same port
-	_, err = manager.CreateInstance("new-port-test", options3)
-	if err != nil {
-		t.Errorf("Expected to reuse port after deletion, got error: %v", err)
-	}
-}
-
-func TestInstanceOperations(t *testing.T) {
-	manager := createTestManager()
-
-	options := &instance.CreateInstanceOptions{
-		LlamaServerOptions: llamacpp.LlamaServerOptions{
-			Model: "/path/to/model.gguf",
-		},
-	}
-
-	// Create instance
-	created, err := manager.CreateInstance("test-instance", options)
-	if err != nil {
-		t.Fatalf("CreateInstance failed: %v", err)
-	}
-
-	// Get instance
-	retrieved, err := manager.GetInstance("test-instance")
-	if err != nil {
-		t.Fatalf("GetInstance failed: %v", err)
-	}
-	if retrieved.Name != created.Name {
-		t.Errorf("Expected name %q, got %q", created.Name, retrieved.Name)
-	}
-
-	// Update instance
-	newOptions := &instance.CreateInstanceOptions{
-		LlamaServerOptions: llamacpp.LlamaServerOptions{
-			Model: "/path/to/new-model.gguf",
-			Port:  8081,
-		},
-	}
-
-	updated, err := manager.UpdateInstance("test-instance", newOptions)
-	if err != nil {
-		t.Fatalf("UpdateInstance failed: %v", err)
-	}
-	if updated.GetOptions().Model != "/path/to/new-model.gguf" {
-		t.Errorf("Expected model '/path/to/new-model.gguf', got %q", updated.GetOptions().Model)
-	}
-
-	// List instances
-	instances, err := manager.ListInstances()
-	if err != nil {
-		t.Fatalf("ListInstances failed: %v", err)
-	}
-	if len(instances) != 1 {
-		t.Errorf("Expected 1 instance, got %d", len(instances))
-	}
-
-	// Delete instance
-	err = manager.DeleteInstance("test-instance")
-	if err != nil {
-		t.Fatalf("DeleteInstance failed: %v", err)
-	}
-
-	_, err = manager.GetInstance("test-instance")
-	if err == nil {
-		t.Error("Instance should not exist after deletion")
-	}
-
-	// Test operations on non-existent instances
-	_, err = manager.GetInstance("nonexistent")
-	if err == nil || !strings.Contains(err.Error(), "not found") {
-		t.Errorf("Expected 'not found' error, got: %v", err)
-	}
-
-	err = manager.DeleteInstance("nonexistent")
-	if err == nil || !strings.Contains(err.Error(), "not found") {
-		t.Errorf("Expected 'not found' error, got: %v", err)
-	}
-
-	_, err = manager.UpdateInstance("nonexistent", options)
-	if err == nil || !strings.Contains(err.Error(), "not found") {
-		t.Errorf("Expected 'not found' error, got: %v", err)
-	}
-}
-
 func TestPersistence(t *testing.T) {
 	tempDir := t.TempDir()
 
+	backendConfig := config.BackendConfig{
+		LlamaCpp: config.BackendSettings{
+			Command: "llama-server",
+		},
+		MLX: config.BackendSettings{
+			Command: "mlx_lm.server",
+		},
+	}
+
 	cfg := config.InstancesConfig{
 		PortRange:            [2]int{8000, 9000},
 		InstancesDir:         tempDir,
@@ -263,9 +69,10 @@ func TestPersistence(t *testing.T) {
 	}
 
 	// Test instance persistence on creation
-	manager1 := manager.NewInstanceManager(cfg)
+	manager1 := manager.NewInstanceManager(backendConfig, cfg)
 	options := &instance.CreateInstanceOptions{
-		LlamaServerOptions: llamacpp.LlamaServerOptions{
+		BackendType: backends.BackendTypeLlamaCpp,
+		LlamaServerOptions: &llamacpp.LlamaServerOptions{
 			Model: "/path/to/model.gguf",
 			Port:  8080,
 		},
@@ -283,7 +90,7 @@ func TestPersistence(t *testing.T) {
 	}
 
 	// Test loading instances from disk
-	manager2 := manager.NewInstanceManager(cfg)
+	manager2 := manager.NewInstanceManager(backendConfig, cfg)
 	instances, err := manager2.ListInstances()
 	if err != nil {
 		t.Fatalf("ListInstances failed: %v", err)
@@ -312,119 +119,27 @@ func TestPersistence(t *testing.T) {
 	}
 }
 
-func TestTimeoutFunctionality(t *testing.T) {
-	// Test timeout checker initialization
-	cfg := config.InstancesConfig{
-		PortRange:            [2]int{8000, 9000},
-		TimeoutCheckInterval: 10,
-		MaxInstances:         5,
-	}
-
-	manager := manager.NewInstanceManager(cfg)
-	if manager == nil {
-		t.Fatal("Manager should be initialized with timeout checker")
-	}
-	manager.Shutdown() // Clean up
-
-	// Test timeout configuration and logic without starting the actual process
-	testManager := createTestManager()
-	defer testManager.Shutdown()
-
-	idleTimeout := 1 // 1 minute
-	options := &instance.CreateInstanceOptions{
-		IdleTimeout: &idleTimeout,
-		LlamaServerOptions: llamacpp.LlamaServerOptions{
-			Model: "/path/to/model.gguf",
-		},
-	}
-
-	inst, err := testManager.CreateInstance("timeout-test", options)
-	if err != nil {
-		t.Fatalf("CreateInstance failed: %v", err)
-	}
-
-	// Test timeout configuration is properly set
-	if inst.GetOptions().IdleTimeout == nil {
-		t.Fatal("Instance should have idle timeout configured")
-	}
-	if *inst.GetOptions().IdleTimeout != 1 {
-		t.Errorf("Expected idle timeout 1 minute, got %d", *inst.GetOptions().IdleTimeout)
-	}
-
-	// Test timeout logic without actually starting the process
-	// Create a mock time provider to simulate timeout
-	mockTime := NewMockTimeProvider(time.Now())
-	inst.SetTimeProvider(mockTime)
-
-	// Set instance to running state so timeout logic can work
-	inst.Running = true
-
-	// Simulate instance being "running" for timeout check (without actual process)
-	// We'll test the ShouldTimeout logic directly
-	inst.UpdateLastRequestTime()
-
-	// Initially should not timeout (just updated)
-	if inst.ShouldTimeout() {
-		t.Error("Instance should not timeout immediately after request")
-	}
-
-	// Advance time to trigger timeout
-	mockTime.SetTime(time.Now().Add(2 * time.Minute))
-
-	// Now it should timeout
-	if !inst.ShouldTimeout() {
-		t.Error("Instance should timeout after idle period")
-	}
-
-	// Reset running state to avoid shutdown issues
-	inst.Running = false
-
-	// Test that instance without timeout doesn't timeout
-	noTimeoutOptions := &instance.CreateInstanceOptions{
-		LlamaServerOptions: llamacpp.LlamaServerOptions{
-			Model: "/path/to/model.gguf",
-		},
-		// No IdleTimeout set
-	}
-
-	noTimeoutInst, err := testManager.CreateInstance("no-timeout-test", noTimeoutOptions)
-	if err != nil {
-		t.Fatalf("CreateInstance failed: %v", err)
-	}
-
-	noTimeoutInst.SetTimeProvider(mockTime)
-	noTimeoutInst.Running = true // Set to running for timeout check
-	noTimeoutInst.UpdateLastRequestTime()
-
-	// Even with time advanced, should not timeout
-	if noTimeoutInst.ShouldTimeout() {
-		t.Error("Instance without timeout configuration should never timeout")
-	}
-
-	// Reset running state to avoid shutdown issues
-	noTimeoutInst.Running = false
-}
-
 func TestConcurrentAccess(t *testing.T) {
-	manager := createTestManager()
-	defer manager.Shutdown()
+	mgr := createTestManager()
+	defer mgr.Shutdown()
 
 	// Test concurrent operations
 	var wg sync.WaitGroup
 	errChan := make(chan error, 10)
 
 	// Concurrent instance creation
-	for i := 0; i < 5; i++ {
+	for i := range 5 {
 		wg.Add(1)
 		go func(index int) {
 			defer wg.Done()
 			options := &instance.CreateInstanceOptions{
-				LlamaServerOptions: llamacpp.LlamaServerOptions{
+				BackendType: backends.BackendTypeLlamaCpp,
+				LlamaServerOptions: &llamacpp.LlamaServerOptions{
 					Model: "/path/to/model.gguf",
 				},
 			}
 			instanceName := fmt.Sprintf("concurrent-test-%d", index)
-			if _, err := manager.CreateInstance(instanceName, options); err != nil {
+			if _, err := mgr.CreateInstance(instanceName, options); err != nil {
 				errChan <- err
 			}
 		}(i)
@@ -435,7 +150,7 @@ func TestConcurrentAccess(t *testing.T) {
 		wg.Add(1)
 		go func() {
 			defer wg.Done()
-			if _, err := manager.ListInstances(); err != nil {
+			if _, err := mgr.ListInstances(); err != nil {
 				errChan <- err
 			}
 		}()
@@ -451,59 +166,109 @@ func TestConcurrentAccess(t *testing.T) {
 }
 
 func TestShutdown(t *testing.T) {
-	manager := createTestManager()
+	mgr := createTestManager()
 
 	// Create test instance
 	options := &instance.CreateInstanceOptions{
-		LlamaServerOptions: llamacpp.LlamaServerOptions{
+		BackendType: backends.BackendTypeLlamaCpp,
+		LlamaServerOptions: &llamacpp.LlamaServerOptions{
 			Model: "/path/to/model.gguf",
 		},
 	}
-	_, err := manager.CreateInstance("test-instance", options)
+	_, err := mgr.CreateInstance("test-instance", options)
 	if err != nil {
 		t.Fatalf("CreateInstance failed: %v", err)
 	}
 
 	// Shutdown should not panic
-	manager.Shutdown()
+	mgr.Shutdown()
 
 	// Multiple shutdowns should not panic
-	manager.Shutdown()
+	mgr.Shutdown()
 }
 
 // Helper function to create a test manager with standard config
 func createTestManager() manager.InstanceManager {
+	backendConfig := config.BackendConfig{
+		LlamaCpp: config.BackendSettings{
+			Command: "llama-server",
+		},
+		MLX: config.BackendSettings{
+			Command: "mlx_lm.server",
+		},
+	}
+
 	cfg := config.InstancesConfig{
 		PortRange:            [2]int{8000, 9000},
 		LogsDir:              "/tmp/test",
 		MaxInstances:         10,
-		LlamaExecutable:      "llama-server",
 		DefaultAutoRestart:   true,
 		DefaultMaxRestarts:   3,
 		DefaultRestartDelay:  5,
 		TimeoutCheckInterval: 5,
 	}
-	return manager.NewInstanceManager(cfg)
+	return manager.NewInstanceManager(backendConfig, cfg)
 }
 
-// Helper for timeout tests
-type MockTimeProvider struct {
-	currentTime time.Time
-	mu          sync.RWMutex
-}
+func TestAutoRestartDisabledInstanceStatus(t *testing.T) {
+	tempDir := t.TempDir()
 
-func NewMockTimeProvider(t time.Time) *MockTimeProvider {
-	return &MockTimeProvider{currentTime: t}
-}
+	backendConfig := config.BackendConfig{
+		LlamaCpp: config.BackendSettings{
+			Command: "llama-server",
+		},
+	}
 
-func (m *MockTimeProvider) Now() time.Time {
-	m.mu.RLock()
-	defer m.mu.RUnlock()
-	return m.currentTime
-}
+	cfg := config.InstancesConfig{
+		PortRange:            [2]int{8000, 9000},
+		InstancesDir:         tempDir,
+		MaxInstances:         10,
+		TimeoutCheckInterval: 5,
+	}
 
-func (m *MockTimeProvider) SetTime(t time.Time) {
-	m.mu.Lock()
-	defer m.mu.Unlock()
-	m.currentTime = t
+	// Create first manager and instance with auto-restart disabled
+	manager1 := manager.NewInstanceManager(backendConfig, cfg)
+
+	autoRestart := false
+	options := &instance.CreateInstanceOptions{
+		BackendType: backends.BackendTypeLlamaCpp,
+		AutoRestart: &autoRestart,
+		LlamaServerOptions: &llamacpp.LlamaServerOptions{
+			Model: "/path/to/model.gguf",
+			Port:  8080,
+		},
+	}
+
+	inst, err := manager1.CreateInstance("test-instance", options)
+	if err != nil {
+		t.Fatalf("CreateInstance failed: %v", err)
+	}
+
+	// Simulate instance being in running state when persisted
+	// (this would happen if the instance was running when llamactl was stopped)
+	inst.SetStatus(instance.Running)
+
+	// Shutdown first manager
+	manager1.Shutdown()
+
+	// Create second manager (simulating restart of llamactl)
+	manager2 := manager.NewInstanceManager(backendConfig, cfg)
+
+	// Get the loaded instance
+	loadedInst, err := manager2.GetInstance("test-instance")
+	if err != nil {
+		t.Fatalf("GetInstance failed: %v", err)
+	}
+
+	// The instance should be marked as Stopped, not Running
+	// because auto-restart is disabled
+	if loadedInst.IsRunning() {
+		t.Errorf("Expected instance with auto-restart disabled to be stopped after manager restart, but it was running")
+	}
+
+	if loadedInst.GetStatus() != instance.Stopped {
+		t.Errorf("Expected instance status to be Stopped, got %v", loadedInst.GetStatus())
+	}
+
+	manager2.Shutdown()
 }

pkg/manager/operations.go

@@ -2,12 +2,15 @@ package manager
 
 import (
 	"fmt"
+	"llamactl/pkg/backends"
 	"llamactl/pkg/instance"
 	"llamactl/pkg/validation"
 	"os"
 	"path/filepath"
 )
 
+type MaxRunningInstancesError error
+
 // ListInstances returns a list of all instances managed by the instance manager.
 func (im *instanceManager) ListInstances() ([]*instance.Process, error) {
 	im.mu.RLock()
@@ -50,24 +53,17 @@ func (im *instanceManager) CreateInstance(name string, options *instance.CreateI
 		return nil, fmt.Errorf("instance with name %s already exists", name)
 	}
 
-	// Assign a port if not specified
-	if options.Port == 0 {
-		port, err := im.getNextAvailablePort()
-		if err != nil {
-			return nil, fmt.Errorf("failed to get next available port: %w", err)
-		}
-		options.Port = port
-	} else {
-		// Validate the specified port
-		if _, exists := im.ports[options.Port]; exists {
-			return nil, fmt.Errorf("port %d is already in use", options.Port)
-		}
-		im.ports[options.Port] = true
+	// Assign and validate port for backend-specific options
+	if err := im.assignAndValidatePort(options); err != nil {
+		return nil, err
 	}
 
-	inst := instance.NewInstance(name, &im.instancesConfig, options)
+	statusCallback := func(oldStatus, newStatus instance.InstanceStatus) {
+		im.onStatusChange(name, oldStatus, newStatus)
+	}
+
+	inst := instance.NewInstance(name, &im.backendsConfig, &im.instancesConfig, options, statusCallback)
 	im.instances[inst.Name] = inst
-	im.ports[options.Port] = true
 
 	if err := im.persistInstance(inst); err != nil {
 		return nil, fmt.Errorf("failed to persist instance %s: %w", name, err)
@@ -109,7 +105,7 @@ func (im *instanceManager) UpdateInstance(name string, options *instance.CreateI
 	}
 
 	// Check if instance is running before updating options
-	wasRunning := instance.Running
+	wasRunning := instance.IsRunning()
 
 	// If the instance is running, stop it first
 	if wasRunning {
@@ -147,11 +143,11 @@ func (im *instanceManager) DeleteInstance(name string) error {
 		return fmt.Errorf("instance with name %s not found", name)
 	}
 
-	if instance.Running {
+	if instance.IsRunning() {
 		return fmt.Errorf("instance with name %s is still running, stop it before deleting", name)
 	}
 
-	delete(im.ports, instance.GetOptions().Port)
+	delete(im.ports, instance.GetPort())
 	delete(im.instances, name)
 
 	// Delete the instance's config file if persistence is enabled
@@ -168,15 +164,20 @@ func (im *instanceManager) DeleteInstance(name string) error {
 func (im *instanceManager) StartInstance(name string) (*instance.Process, error) {
 	im.mu.RLock()
 	instance, exists := im.instances[name]
+	maxRunningExceeded := len(im.runningInstances) >= im.instancesConfig.MaxRunningInstances && im.instancesConfig.MaxRunningInstances != -1
 	im.mu.RUnlock()
 
 	if !exists {
 		return nil, fmt.Errorf("instance with name %s not found", name)
 	}
-	if instance.Running {
+	if instance.IsRunning() {
 		return instance, fmt.Errorf("instance with name %s is already running", name)
 	}
 
+	if maxRunningExceeded {
+		return nil, MaxRunningInstancesError(fmt.Errorf("maximum number of running instances (%d) reached", im.instancesConfig.MaxRunningInstances))
+	}
+
 	if err := instance.Start(); err != nil {
 		return nil, fmt.Errorf("failed to start instance %s: %w", name, err)
 	}
@@ -191,6 +192,17 @@ func (im *instanceManager) StartInstance(name string) (*instance.Process, error)
 	return instance, nil
 }
 
+func (im *instanceManager) IsMaxRunningInstancesReached() bool {
+	im.mu.RLock()
+	defer im.mu.RUnlock()
+
+	if im.instancesConfig.MaxRunningInstances != -1 && len(im.runningInstances) >= im.instancesConfig.MaxRunningInstances {
+		return true
+	}
+
+	return false
+}
+
 // StopInstance stops a running instance and returns it.
 func (im *instanceManager) StopInstance(name string) (*instance.Process, error) {
 	im.mu.RLock()
@@ -200,7 +212,7 @@ func (im *instanceManager) StopInstance(name string) (*instance.Process, error)
 	if !exists {
 		return nil, fmt.Errorf("instance with name %s not found", name)
 	}
-	if !instance.Running {
+	if !instance.IsRunning() {
 		return instance, fmt.Errorf("instance with name %s is already stopped", name)
 	}
 
@@ -240,3 +252,65 @@ func (im *instanceManager) GetInstanceLogs(name string) (string, error) {
 	// TODO: Implement actual log retrieval logic
 	return fmt.Sprintf("Logs for instance %s", name), nil
 }
+
+// getPortFromOptions extracts the port from backend-specific options
+func (im *instanceManager) getPortFromOptions(options *instance.CreateInstanceOptions) int {
+	switch options.BackendType {
+	case backends.BackendTypeLlamaCpp:
+		if options.LlamaServerOptions != nil {
+			return options.LlamaServerOptions.Port
+		}
+	case backends.BackendTypeMlxLm:
+		if options.MlxServerOptions != nil {
+			return options.MlxServerOptions.Port
+		}
+	case backends.BackendTypeVllm:
+		if options.VllmServerOptions != nil {
+			return options.VllmServerOptions.Port
+		}
+	}
+	return 0
+}
+
+// setPortInOptions sets the port in backend-specific options
+func (im *instanceManager) setPortInOptions(options *instance.CreateInstanceOptions, port int) {
+	switch options.BackendType {
+	case backends.BackendTypeLlamaCpp:
+		if options.LlamaServerOptions != nil {
+			options.LlamaServerOptions.Port = port
+		}
+	case backends.BackendTypeMlxLm:
+		if options.MlxServerOptions != nil {
+			options.MlxServerOptions.Port = port
+		}
+	case backends.BackendTypeVllm:
+		if options.VllmServerOptions != nil {
+			options.VllmServerOptions.Port = port
+		}
+	}
+}
+
+// assignAndValidatePort assigns a port if not specified and validates it's not in use
+func (im *instanceManager) assignAndValidatePort(options *instance.CreateInstanceOptions) error {
+	currentPort := im.getPortFromOptions(options)
+
+	if currentPort == 0 {
+		// Assign a port if not specified
+		port, err := im.getNextAvailablePort()
+		if err != nil {
+			return fmt.Errorf("failed to get next available port: %w", err)
+		}
+		im.setPortInOptions(options, port)
+		// Mark the port as used
+		im.ports[port] = true
+	} else {
+		// Validate the specified port
+		if _, exists := im.ports[currentPort]; exists {
+			return fmt.Errorf("port %d is already in use", currentPort)
+		}
+		// Mark the port as used
+		im.ports[currentPort] = true
+	}
+
+	return nil
+}

pkg/manager/operations_test.go

@@ -0,0 +1,237 @@
+package manager_test
+
+import (
+	"llamactl/pkg/backends"
+	"llamactl/pkg/backends/llamacpp"
+	"llamactl/pkg/config"
+	"llamactl/pkg/instance"
+	"llamactl/pkg/manager"
+	"strings"
+	"testing"
+)
+
+func TestCreateInstance_Success(t *testing.T) {
+	manager := createTestManager()
+
+	options := &instance.CreateInstanceOptions{
+		BackendType: backends.BackendTypeLlamaCpp,
+		LlamaServerOptions: &llamacpp.LlamaServerOptions{
+			Model: "/path/to/model.gguf",
+			Port:  8080,
+		},
+	}
+
+	inst, err := manager.CreateInstance("test-instance", options)
+	if err != nil {
+		t.Fatalf("CreateInstance failed: %v", err)
+	}
+
+	if inst.Name != "test-instance" {
+		t.Errorf("Expected instance name 'test-instance', got %q", inst.Name)
+	}
+	if inst.GetStatus() != instance.Stopped {
+		t.Error("New instance should not be running")
+	}
+	if inst.GetPort() != 8080 {
+		t.Errorf("Expected port 8080, got %d", inst.GetPort())
+	}
+}
+
+func TestCreateInstance_ValidationAndLimits(t *testing.T) {
+	// Test duplicate names
+	mngr := createTestManager()
+	options := &instance.CreateInstanceOptions{
+		BackendType: backends.BackendTypeLlamaCpp,
+		LlamaServerOptions: &llamacpp.LlamaServerOptions{
+			Model: "/path/to/model.gguf",
+		},
+	}
+
+	_, err := mngr.CreateInstance("test-instance", options)
+	if err != nil {
+		t.Fatalf("First CreateInstance failed: %v", err)
+	}
+
+	// Try to create duplicate
+	_, err = mngr.CreateInstance("test-instance", options)
+	if err == nil {
+		t.Error("Expected error for duplicate instance name")
+	}
+	if !strings.Contains(err.Error(), "already exists") {
+		t.Errorf("Expected duplicate name error, got: %v", err)
+	}
+
+	// Test max instances limit
+	backendConfig := config.BackendConfig{
+		LlamaCpp: config.BackendSettings{
+			Command: "llama-server",
+		},
+		MLX: config.BackendSettings{
+			Command: "mlx_lm.server",
+		},
+	}
+	cfg := config.InstancesConfig{
+		PortRange:            [2]int{8000, 9000},
+		MaxInstances:         1, // Very low limit for testing
+		TimeoutCheckInterval: 5,
+	}
+	limitedManager := manager.NewInstanceManager(backendConfig, cfg)
+
+	_, err = limitedManager.CreateInstance("instance1", options)
+	if err != nil {
+		t.Fatalf("CreateInstance 1 failed: %v", err)
+	}
+
+	// This should fail due to max instances limit
+	_, err = limitedManager.CreateInstance("instance2", options)
+	if err == nil {
+		t.Error("Expected error when exceeding max instances limit")
+	}
+	if !strings.Contains(err.Error(), "maximum number of instances") {
+		t.Errorf("Expected max instances error, got: %v", err)
+	}
+}
+
+func TestPortManagement(t *testing.T) {
+	manager := createTestManager()
+
+	// Test auto port assignment
+	options1 := &instance.CreateInstanceOptions{
+		BackendType: backends.BackendTypeLlamaCpp,
+		LlamaServerOptions: &llamacpp.LlamaServerOptions{
+			Model: "/path/to/model.gguf",
+		},
+	}
+
+	inst1, err := manager.CreateInstance("instance1", options1)
+	if err != nil {
+		t.Fatalf("CreateInstance failed: %v", err)
+	}
+
+	port1 := inst1.GetPort()
+	if port1 < 8000 || port1 > 9000 {
+		t.Errorf("Expected port in range 8000-9000, got %d", port1)
+	}
+
+	// Test port conflict detection
+	options2 := &instance.CreateInstanceOptions{
+		BackendType: backends.BackendTypeLlamaCpp,
+		LlamaServerOptions: &llamacpp.LlamaServerOptions{
+			Model: "/path/to/model2.gguf",
+			Port:  port1, // Same port - should conflict
+		},
+	}
+
+	_, err = manager.CreateInstance("instance2", options2)
+	if err == nil {
+		t.Error("Expected error for port conflict")
+	}
+	if !strings.Contains(err.Error(), "port") && !strings.Contains(err.Error(), "in use") {
+		t.Errorf("Expected port conflict error, got: %v", err)
+	}
+
+	// Test port release on deletion
+	specificPort := 8080
+	options3 := &instance.CreateInstanceOptions{
+		BackendType: backends.BackendTypeLlamaCpp,
+		LlamaServerOptions: &llamacpp.LlamaServerOptions{
+			Model: "/path/to/model.gguf",
+			Port:  specificPort,
+		},
+	}
+
+	_, err = manager.CreateInstance("port-test", options3)
+	if err != nil {
+		t.Fatalf("CreateInstance failed: %v", err)
+	}
+
+	err = manager.DeleteInstance("port-test")
+	if err != nil {
+		t.Fatalf("DeleteInstance failed: %v", err)
+	}
+
+	// Should be able to create new instance with same port
+	_, err = manager.CreateInstance("new-port-test", options3)
+	if err != nil {
+		t.Errorf("Expected to reuse port after deletion, got error: %v", err)
+	}
+}
+
+func TestInstanceOperations(t *testing.T) {
+	manager := createTestManager()
+
+	options := &instance.CreateInstanceOptions{
+		BackendType: backends.BackendTypeLlamaCpp,
+		LlamaServerOptions: &llamacpp.LlamaServerOptions{
+			Model: "/path/to/model.gguf",
+		},
+	}
+
+	// Create instance
+	created, err := manager.CreateInstance("test-instance", options)
+	if err != nil {
+		t.Fatalf("CreateInstance failed: %v", err)
+	}
+
+	// Get instance
+	retrieved, err := manager.GetInstance("test-instance")
+	if err != nil {
+		t.Fatalf("GetInstance failed: %v", err)
+	}
+	if retrieved.Name != created.Name {
+		t.Errorf("Expected name %q, got %q", created.Name, retrieved.Name)
+	}
+
+	// Update instance
+	newOptions := &instance.CreateInstanceOptions{
+		BackendType: backends.BackendTypeLlamaCpp,
+		LlamaServerOptions: &llamacpp.LlamaServerOptions{
+			Model: "/path/to/new-model.gguf",
+			Port:  8081,
+		},
+	}
+
+	updated, err := manager.UpdateInstance("test-instance", newOptions)
+	if err != nil {
+		t.Fatalf("UpdateInstance failed: %v", err)
+	}
+	if updated.GetOptions().LlamaServerOptions.Model != "/path/to/new-model.gguf" {
+		t.Errorf("Expected model '/path/to/new-model.gguf', got %q", updated.GetOptions().LlamaServerOptions.Model)
+	}
+
+	// List instances
+	instances, err := manager.ListInstances()
+	if err != nil {
+		t.Fatalf("ListInstances failed: %v", err)
+	}
+	if len(instances) != 1 {
+		t.Errorf("Expected 1 instance, got %d", len(instances))
+	}
+
+	// Delete instance
+	err = manager.DeleteInstance("test-instance")
+	if err != nil {
+		t.Fatalf("DeleteInstance failed: %v", err)
+	}
+
+	_, err = manager.GetInstance("test-instance")
+	if err == nil {
+		t.Error("Instance should not exist after deletion")
+	}
+
+	// Test operations on non-existent instances
+	_, err = manager.GetInstance("nonexistent")
+	if err == nil || !strings.Contains(err.Error(), "not found") {
+		t.Errorf("Expected 'not found' error, got: %v", err)
+	}
+
+	err = manager.DeleteInstance("nonexistent")
+	if err == nil || !strings.Contains(err.Error(), "not found") {
+		t.Errorf("Expected 'not found' error, got: %v", err)
+	}
+
+	_, err = manager.UpdateInstance("nonexistent", options)
+	if err == nil || !strings.Contains(err.Error(), "not found") {
+		t.Errorf("Expected 'not found' error, got: %v", err)
+	}
+}

pkg/manager/timeout.go

@@ -1,6 +1,10 @@
 package manager
 
-import "log"
+import (
+	"fmt"
+	"llamactl/pkg/instance"
+	"log"
+)
 
 func (im *instanceManager) checkAllTimeouts() {
 	im.mu.RLock()
@@ -24,3 +28,37 @@ func (im *instanceManager) checkAllTimeouts() {
 		}
 	}
 }
+
+// EvictLRUInstance finds and stops the least recently used running instance.
+func (im *instanceManager) EvictLRUInstance() error {
+	im.mu.RLock()
+	var lruInstance *instance.Process
+
+	for name := range im.runningInstances {
+		inst := im.instances[name]
+		if inst == nil {
+			continue
+		}
+
+		if inst.GetOptions() != nil && inst.GetOptions().IdleTimeout != nil && *inst.GetOptions().IdleTimeout <= 0 {
+			continue // Skip instances without idle timeout
+		}
+
+		if lruInstance == nil {
+			lruInstance = inst
+		}
+
+		if inst.LastRequestTime() < lruInstance.LastRequestTime() {
+			lruInstance = inst
+		}
+	}
+	im.mu.RUnlock()
+
+	if lruInstance == nil {
+		return fmt.Errorf("failed to find lru instance")
+	}
+
+	// Evict Instance
+	_, err := im.StopInstance(lruInstance.Name)
+	return err
+}

pkg/manager/timeout_test.go

@@ -0,0 +1,332 @@
+package manager_test
+
+import (
+	"llamactl/pkg/backends"
+	"llamactl/pkg/backends/llamacpp"
+	"llamactl/pkg/config"
+	"llamactl/pkg/instance"
+	"llamactl/pkg/manager"
+	"sync"
+	"testing"
+	"time"
+)
+
+func TestTimeoutFunctionality(t *testing.T) {
+	// Test timeout checker initialization
+	backendConfig := config.BackendConfig{
+		LlamaCpp: config.BackendSettings{Command: "llama-server"},
+		MLX:      config.BackendSettings{Command: "mlx_lm.server"},
+	}
+	cfg := config.InstancesConfig{
+		PortRange:            [2]int{8000, 9000},
+		TimeoutCheckInterval: 10,
+		MaxInstances:         5,
+	}
+
+	manager := manager.NewInstanceManager(backendConfig, cfg)
+	if manager == nil {
+		t.Fatal("Manager should be initialized with timeout checker")
+	}
+	manager.Shutdown() // Clean up
+
+	// Test timeout configuration and logic without starting the actual process
+	testManager := createTestManager()
+	defer testManager.Shutdown()
+
+	idleTimeout := 1 // 1 minute
+	options := &instance.CreateInstanceOptions{
+		IdleTimeout: &idleTimeout,
+		BackendType: backends.BackendTypeLlamaCpp,
+		LlamaServerOptions: &llamacpp.LlamaServerOptions{
+			Model: "/path/to/model.gguf",
+		},
+	}
+
+	inst, err := testManager.CreateInstance("timeout-test", options)
+	if err != nil {
+		t.Fatalf("CreateInstance failed: %v", err)
+	}
+
+	// Test timeout configuration is properly set
+	if inst.GetOptions().IdleTimeout == nil {
+		t.Fatal("Instance should have idle timeout configured")
+	}
+	if *inst.GetOptions().IdleTimeout != 1 {
+		t.Errorf("Expected idle timeout 1 minute, got %d", *inst.GetOptions().IdleTimeout)
+	}
+
+	// Test timeout logic without actually starting the process
+	// Create a mock time provider to simulate timeout
+	mockTime := NewMockTimeProvider(time.Now())
+	inst.SetTimeProvider(mockTime)
+
+	// Set instance to running state so timeout logic can work
+	inst.SetStatus(instance.Running)
+
+	// Simulate instance being "running" for timeout check (without actual process)
+	// We'll test the ShouldTimeout logic directly
+	inst.UpdateLastRequestTime()
+
+	// Initially should not timeout (just updated)
+	if inst.ShouldTimeout() {
+		t.Error("Instance should not timeout immediately after request")
+	}
+
+	// Advance time to trigger timeout
+	mockTime.SetTime(time.Now().Add(2 * time.Minute))
+
+	// Now it should timeout
+	if !inst.ShouldTimeout() {
+		t.Error("Instance should timeout after idle period")
+	}
+
+	// Reset running state to avoid shutdown issues
+	inst.SetStatus(instance.Stopped)
+
+	// Test that instance without timeout doesn't timeout
+	noTimeoutOptions := &instance.CreateInstanceOptions{
+		BackendType: backends.BackendTypeLlamaCpp,
+		LlamaServerOptions: &llamacpp.LlamaServerOptions{
+			Model: "/path/to/model.gguf",
+		},
+		// No IdleTimeout set
+	}
+
+	noTimeoutInst, err := testManager.CreateInstance("no-timeout-test", noTimeoutOptions)
+	if err != nil {
+		t.Fatalf("CreateInstance failed: %v", err)
+	}
+
+	noTimeoutInst.SetTimeProvider(mockTime)
+	noTimeoutInst.SetStatus(instance.Running) // Set to running for timeout check
+	noTimeoutInst.UpdateLastRequestTime()
+
+	// Even with time advanced, should not timeout
+	if noTimeoutInst.ShouldTimeout() {
+		t.Error("Instance without timeout configuration should never timeout")
+	}
+
+	// Reset running state to avoid shutdown issues
+	noTimeoutInst.SetStatus(instance.Stopped)
+}
+
+func TestEvictLRUInstance_Success(t *testing.T) {
+	manager := createTestManager()
+	// Don't defer manager.Shutdown() - we'll handle cleanup manually
+
+	// Create 3 instances with idle timeout enabled (value doesn't matter for LRU logic)
+	options1 := &instance.CreateInstanceOptions{
+		BackendType: backends.BackendTypeLlamaCpp,
+		LlamaServerOptions: &llamacpp.LlamaServerOptions{
+			Model: "/path/to/model1.gguf",
+		},
+		IdleTimeout: func() *int { timeout := 1; return &timeout }(), // Any value > 0
+	}
+	options2 := &instance.CreateInstanceOptions{
+		BackendType: backends.BackendTypeLlamaCpp,
+		LlamaServerOptions: &llamacpp.LlamaServerOptions{
+			Model: "/path/to/model2.gguf",
+		},
+		IdleTimeout: func() *int { timeout := 1; return &timeout }(), // Any value > 0
+	}
+	options3 := &instance.CreateInstanceOptions{
+		BackendType: backends.BackendTypeLlamaCpp,
+		LlamaServerOptions: &llamacpp.LlamaServerOptions{
+			Model: "/path/to/model3.gguf",
+		},
+		IdleTimeout: func() *int { timeout := 1; return &timeout }(), // Any value > 0
+	}
+
+	inst1, err := manager.CreateInstance("instance-1", options1)
+	if err != nil {
+		t.Fatalf("CreateInstance failed: %v", err)
+	}
+	inst2, err := manager.CreateInstance("instance-2", options2)
+	if err != nil {
+		t.Fatalf("CreateInstance failed: %v", err)
+	}
+	inst3, err := manager.CreateInstance("instance-3", options3)
+	if err != nil {
+		t.Fatalf("CreateInstance failed: %v", err)
+	}
+
+	// Set up mock time and set instances to running
+	mockTime := NewMockTimeProvider(time.Now())
+	inst1.SetTimeProvider(mockTime)
+	inst2.SetTimeProvider(mockTime)
+	inst3.SetTimeProvider(mockTime)
+
+	inst1.SetStatus(instance.Running)
+	inst2.SetStatus(instance.Running)
+	inst3.SetStatus(instance.Running)
+
+	// Set different last request times (oldest to newest)
+	// inst1: oldest (will be evicted)
+	inst1.UpdateLastRequestTime()
+
+	mockTime.SetTime(mockTime.Now().Add(1 * time.Minute))
+	inst2.UpdateLastRequestTime()
+
+	mockTime.SetTime(mockTime.Now().Add(1 * time.Minute))
+	inst3.UpdateLastRequestTime()
+
+	// Evict LRU instance (should be inst1)
+	err = manager.EvictLRUInstance()
+	if err != nil {
+		t.Fatalf("EvictLRUInstance failed: %v", err)
+	}
+
+	// Verify inst1 is stopped
+	if inst1.IsRunning() {
+		t.Error("Expected instance-1 to be stopped after eviction")
+	}
+
+	// Verify inst2 and inst3 are still running
+	if !inst2.IsRunning() {
+		t.Error("Expected instance-2 to still be running")
+	}
+	if !inst3.IsRunning() {
+		t.Error("Expected instance-3 to still be running")
+	}
+
+	// Clean up manually - set all to stopped and then shutdown
+	inst2.SetStatus(instance.Stopped)
+	inst3.SetStatus(instance.Stopped)
+}
+
+func TestEvictLRUInstance_NoEligibleInstances(t *testing.T) {
+	// Helper function to create instances with different timeout configurations
+	createInstanceWithTimeout := func(manager manager.InstanceManager, name, model string, timeout *int) *instance.Process {
+		options := &instance.CreateInstanceOptions{
+			BackendType: backends.BackendTypeLlamaCpp,
+			LlamaServerOptions: &llamacpp.LlamaServerOptions{
+				Model: model,
+			},
+			IdleTimeout: timeout,
+		}
+		inst, err := manager.CreateInstance(name, options)
+		if err != nil {
+			t.Fatalf("CreateInstance failed: %v", err)
+		}
+		return inst
+	}
+
+	t.Run("no running instances", func(t *testing.T) {
+		manager := createTestManager()
+		defer manager.Shutdown()
+
+		err := manager.EvictLRUInstance()
+		if err == nil {
+			t.Error("Expected error when no running instances exist")
+		}
+		if err.Error() != "failed to find lru instance" {
+			t.Errorf("Expected 'failed to find lru instance' error, got: %v", err)
+		}
+	})
+
+	t.Run("only instances without timeout", func(t *testing.T) {
+		manager := createTestManager()
+		defer manager.Shutdown()
+
+		// Create instances with various non-eligible timeout configurations
+		zeroTimeout := 0
+		negativeTimeout := -1
+		inst1 := createInstanceWithTimeout(manager, "no-timeout-1", "/path/to/model1.gguf", &zeroTimeout)
+		inst2 := createInstanceWithTimeout(manager, "no-timeout-2", "/path/to/model2.gguf", &negativeTimeout)
+		inst3 := createInstanceWithTimeout(manager, "no-timeout-3", "/path/to/model3.gguf", nil)
+
+		// Set instances to running
+		instances := []*instance.Process{inst1, inst2, inst3}
+		for _, inst := range instances {
+			inst.SetStatus(instance.Running)
+		}
+		defer func() {
+			// Reset instances to stopped to avoid shutdown panics
+			for _, inst := range instances {
+				inst.SetStatus(instance.Stopped)
+			}
+		}()
+
+		// Try to evict - should fail because no eligible instances
+		err := manager.EvictLRUInstance()
+		if err == nil {
+			t.Error("Expected error when no eligible instances exist")
+		}
+		if err.Error() != "failed to find lru instance" {
+			t.Errorf("Expected 'failed to find lru instance' error, got: %v", err)
+		}
+
+		// Verify all instances are still running
+		for i, inst := range instances {
+			if !inst.IsRunning() {
+				t.Errorf("Expected instance %d to still be running", i+1)
+			}
+		}
+	})
+
+	t.Run("mixed instances - evicts only eligible ones", func(t *testing.T) {
+		manager := createTestManager()
+		defer manager.Shutdown()
+
+		// Create mix of instances: some with timeout enabled, some disabled
+		validTimeout := 1
+		zeroTimeout := 0
+		instWithTimeout := createInstanceWithTimeout(manager, "with-timeout", "/path/to/model-with-timeout.gguf", &validTimeout)
+		instNoTimeout1 := createInstanceWithTimeout(manager, "no-timeout-1", "/path/to/model-no-timeout1.gguf", &zeroTimeout)
+		instNoTimeout2 := createInstanceWithTimeout(manager, "no-timeout-2", "/path/to/model-no-timeout2.gguf", nil)
+
+		// Set all instances to running
+		instances := []*instance.Process{instWithTimeout, instNoTimeout1, instNoTimeout2}
+		for _, inst := range instances {
+			inst.SetStatus(instance.Running)
+			inst.UpdateLastRequestTime()
+		}
+		defer func() {
+			// Reset instances to stopped to avoid shutdown panics
+			for _, inst := range instances {
+				if inst.IsRunning() {
+					inst.SetStatus(instance.Stopped)
+				}
+			}
+		}()
+
+		// Evict LRU instance - should only consider the one with timeout
+		err := manager.EvictLRUInstance()
+		if err != nil {
+			t.Fatalf("EvictLRUInstance failed: %v", err)
+		}
+
+		// Verify only the instance with timeout was evicted
+		if instWithTimeout.IsRunning() {
+			t.Error("Expected with-timeout instance to be stopped after eviction")
+		}
+		if !instNoTimeout1.IsRunning() {
+			t.Error("Expected no-timeout-1 instance to still be running")
+		}
+		if !instNoTimeout2.IsRunning() {
+			t.Error("Expected no-timeout-2 instance to still be running")
+		}
+	})
+}
+
+// Helper for timeout tests
+type MockTimeProvider struct {
+	currentTime time.Time
+	mu          sync.RWMutex
+}
+
+func NewMockTimeProvider(t time.Time) *MockTimeProvider {
+	return &MockTimeProvider{currentTime: t}
+}
+
+func (m *MockTimeProvider) Now() time.Time {
+	m.mu.RLock()
+	defer m.mu.RUnlock()
+	return m.currentTime
+}
+
+func (m *MockTimeProvider) SetTime(t time.Time) {
+	m.mu.Lock()
+	defer m.mu.Unlock()
+	m.currentTime = t
+}

pkg/server/handlers.go

@@ -5,6 +5,10 @@ import (
 	"encoding/json"
 	"fmt"
 	"io"
+	"llamactl/pkg/backends"
+	"llamactl/pkg/backends/llamacpp"
+	"llamactl/pkg/backends/mlx"
+	"llamactl/pkg/backends/vllm"
 	"llamactl/pkg/config"
 	"llamactl/pkg/instance"
 	"llamactl/pkg/manager"
@@ -47,12 +51,12 @@ func (h *Handler) VersionHandler() http.HandlerFunc {
 // LlamaServerHelpHandler godoc
 // @Summary Get help for llama server
 // @Description Returns the help text for the llama server command
-// @Tags server
+// @Tags backends
 // @Security ApiKeyAuth
 // @Produces text/plain
 // @Success 200 {string} string "Help text"
 // @Failure 500 {string} string "Internal Server Error"
-// @Router /server/help [get]
+// @Router /backends/llama-cpp/help [get]
 func (h *Handler) LlamaServerHelpHandler() http.HandlerFunc {
 	return func(w http.ResponseWriter, r *http.Request) {
 		helpCmd := exec.Command("llama-server", "--help")
@@ -69,12 +73,12 @@ func (h *Handler) LlamaServerHelpHandler() http.HandlerFunc {
 // LlamaServerVersionHandler godoc
 // @Summary Get version of llama server
 // @Description Returns the version of the llama server command
-// @Tags server
+// @Tags backends
 // @Security ApiKeyAuth
 // @Produces text/plain
 // @Success 200 {string} string "Version information"
 // @Failure 500 {string} string "Internal Server Error"
-// @Router /server/version [get]
+// @Router /backends/llama-cpp/version [get]
 func (h *Handler) LlamaServerVersionHandler() http.HandlerFunc {
 	return func(w http.ResponseWriter, r *http.Request) {
 		versionCmd := exec.Command("llama-server", "--version")
@@ -91,12 +95,12 @@ func (h *Handler) LlamaServerVersionHandler() http.HandlerFunc {
 // LlamaServerListDevicesHandler godoc
 // @Summary List available devices for llama server
 // @Description Returns a list of available devices for the llama server
-// @Tags server
+// @Tags backends
 // @Security ApiKeyAuth
 // @Produces text/plain
 // @Success 200 {string} string "List of devices"
 // @Failure 500 {string} string "Internal Server Error"
-// @Router /server/devices [get]
+// @Router /backends/llama-cpp/devices [get]
 func (h *Handler) LlamaServerListDevicesHandler() http.HandlerFunc {
 	return func(w http.ResponseWriter, r *http.Request) {
 		listCmd := exec.Command("llama-server", "--list-devices")
@@ -127,11 +131,16 @@ func (h *Handler) ListInstances() http.HandlerFunc {
 			return
 		}
 
-		w.Header().Set("Content-Type", "application/json")
-		if err := json.NewEncoder(w).Encode(instances); err != nil {
+		// Marshal to bytes first to set Content-Length header
+		data, err := json.Marshal(instances)
+		if err != nil {
 			http.Error(w, "Failed to encode instances: "+err.Error(), http.StatusInternalServerError)
 			return
 		}
+
+		w.Header().Set("Content-Type", "application/json")
+		w.Header().Set("Content-Length", strconv.Itoa(len(data)))
+		w.Write(data)
 	}
 }
 
@@ -198,7 +207,7 @@ func (h *Handler) GetInstance() http.HandlerFunc {
 
 		inst, err := h.InstanceManager.GetInstance(name)
 		if err != nil {
-			http.Error(w, "Failed to get instance: "+err.Error(), http.StatusInternalServerError)
+			http.Error(w, "Invalid instance: "+err.Error(), http.StatusBadRequest)
 			return
 		}
 
@@ -272,6 +281,12 @@ func (h *Handler) StartInstance() http.HandlerFunc {
 
 		inst, err := h.InstanceManager.StartInstance(name)
 		if err != nil {
+			// Check if error is due to maximum running instances limit
+			if _, ok := err.(manager.MaxRunningInstancesError); ok {
+				http.Error(w, err.Error(), http.StatusConflict)
+				return
+			}
+
 			http.Error(w, "Failed to start instance: "+err.Error(), http.StatusInternalServerError)
 			return
 		}
@@ -451,7 +466,7 @@ func (h *Handler) ProxyToInstance() http.HandlerFunc {
 			return
 		}
 
-		if !inst.Running {
+		if !inst.IsRunning() {
 			http.Error(w, "Instance is not running", http.StatusServiceUnavailable)
 			return
 		}
@@ -465,29 +480,15 @@ func (h *Handler) ProxyToInstance() http.HandlerFunc {
 
 		// Strip the "/api/v1/instances/<name>/proxy" prefix from the request URL
 		prefix := fmt.Sprintf("/api/v1/instances/%s/proxy", name)
-		proxyPath := r.URL.Path[len(prefix):]
-
-		// Ensure the proxy path starts with "/"
-		if !strings.HasPrefix(proxyPath, "/") {
-			proxyPath = "/" + proxyPath
-		}
+		r.URL.Path = strings.TrimPrefix(r.URL.Path, prefix)
 
 		// Update the last request time for the instance
 		inst.UpdateLastRequestTime()
 
-		// Modify the request to remove the proxy prefix
-		originalPath := r.URL.Path
-		r.URL.Path = proxyPath
-
 		// Set forwarded headers
 		r.Header.Set("X-Forwarded-Host", r.Header.Get("Host"))
 		r.Header.Set("X-Forwarded-Proto", "http")
 
-		// Restore original path for logging purposes
-		defer func() {
-			r.URL.Path = originalPath
-		}()
-
 		// Forward the request using the cached proxy
 		proxy.ServeHTTP(w, r)
 	}
@@ -541,7 +542,7 @@ func (h *Handler) OpenAIListInstances() http.HandlerFunc {
 // @Accept json
 // @Produces json
 // @Success 200 "OpenAI response"
-// @Failure 400 {string} string "Invalid request body or model name"
+// @Failure 400 {string} string "Invalid request body or instance name"
 // @Failure 500 {string} string "Internal Server Error"
 // @Router /v1/ [post]
 func (h *Handler) OpenAIProxy() http.HandlerFunc {
@@ -554,7 +555,7 @@ func (h *Handler) OpenAIProxy() http.HandlerFunc {
 		}
 		r.Body.Close()
 
-		// Parse the body to extract model name
+		// Parse the body to extract instance name
 		var requestBody map[string]any
 		if err := json.Unmarshal(bodyBytes, &requestBody); err != nil {
 			http.Error(w, "Invalid request body", http.StatusBadRequest)
@@ -563,35 +564,49 @@ func (h *Handler) OpenAIProxy() http.HandlerFunc {
 
 		modelName, ok := requestBody["model"].(string)
 		if !ok || modelName == "" {
-			http.Error(w, "Model name is required", http.StatusBadRequest)
+			http.Error(w, "Instance name is required", http.StatusBadRequest)
 			return
 		}
 
-		// Route to the appropriate inst based on model name
+		// Route to the appropriate inst based on instance name
 		inst, err := h.InstanceManager.GetInstance(modelName)
 		if err != nil {
-			http.Error(w, "Failed to get instance: "+err.Error(), http.StatusInternalServerError)
+			http.Error(w, "Invalid instance: "+err.Error(), http.StatusBadRequest)
 			return
 		}
 
-		if !inst.Running {
-			if inst.GetOptions().OnDemandStart != nil && *inst.GetOptions().OnDemandStart {
-				// If on-demand start is enabled, start the instance
-				if _, err := h.InstanceManager.StartInstance(modelName); err != nil {
-					http.Error(w, "Failed to start instance: "+err.Error(), http.StatusInternalServerError)
-					return
-				}
-
-				// Wait for the instance to become healthy before proceeding
-				if err := inst.WaitForHealthy(h.cfg.Instances.OnDemandStartTimeout); err != nil { // 2 minutes timeout
-					http.Error(w, "Instance failed to become healthy: "+err.Error(), http.StatusServiceUnavailable)
-					return
-				}
-
-			} else {
+		if !inst.IsRunning() {
+			options := inst.GetOptions()
+			allowOnDemand := options != nil && options.OnDemandStart != nil && *options.OnDemandStart
+			if !allowOnDemand {
 				http.Error(w, "Instance is not running", http.StatusServiceUnavailable)
 				return
 			}
+
+			if h.InstanceManager.IsMaxRunningInstancesReached() {
+				if h.cfg.Instances.EnableLRUEviction {
+					err := h.InstanceManager.EvictLRUInstance()
+					if err != nil {
+						http.Error(w, "Cannot start Instance, failed to evict instance "+err.Error(), http.StatusInternalServerError)
+						return
+					}
+				} else {
+					http.Error(w, "Cannot start Instance, maximum number of instances reached", http.StatusConflict)
+					return
+				}
+			}
+
+			// If on-demand start is enabled, start the instance
+			if _, err := h.InstanceManager.StartInstance(modelName); err != nil {
+				http.Error(w, "Failed to start instance: "+err.Error(), http.StatusInternalServerError)
+				return
+			}
+
+			// Wait for the instance to become healthy before proceeding
+			if err := inst.WaitForHealthy(h.cfg.Instances.OnDemandStartTimeout); err != nil { // 2 minutes timeout
+				http.Error(w, "Instance failed to become healthy: "+err.Error(), http.StatusServiceUnavailable)
+				return
+			}
 		}
 
 		proxy, err := inst.GetProxy()
@@ -610,3 +625,241 @@ func (h *Handler) OpenAIProxy() http.HandlerFunc {
 		proxy.ServeHTTP(w, r)
 	}
 }
+
+func (h *Handler) LlamaCppProxy(onDemandStart bool) http.HandlerFunc {
+	return func(w http.ResponseWriter, r *http.Request) {
+
+		// Get the instance name from the URL parameter
+		name := chi.URLParam(r, "name")
+		if name == "" {
+			http.Error(w, "Instance name cannot be empty", http.StatusBadRequest)
+			return
+		}
+
+		// Route to the appropriate inst based on instance name
+		inst, err := h.InstanceManager.GetInstance(name)
+		if err != nil {
+			http.Error(w, "Invalid instance: "+err.Error(), http.StatusBadRequest)
+			return
+		}
+
+		options := inst.GetOptions()
+		if options == nil {
+			http.Error(w, "Cannot obtain Instance's options", http.StatusInternalServerError)
+			return
+		}
+
+		if options.BackendType != backends.BackendTypeLlamaCpp {
+			http.Error(w, "Instance is not a llama.cpp server.", http.StatusBadRequest)
+			return
+		}
+
+		if !inst.IsRunning() {
+
+			if !(onDemandStart && options.OnDemandStart != nil && *options.OnDemandStart) {
+				http.Error(w, "Instance is not running", http.StatusServiceUnavailable)
+				return
+			}
+
+			if h.InstanceManager.IsMaxRunningInstancesReached() {
+				if h.cfg.Instances.EnableLRUEviction {
+					err := h.InstanceManager.EvictLRUInstance()
+					if err != nil {
+						http.Error(w, "Cannot start Instance, failed to evict instance "+err.Error(), http.StatusInternalServerError)
+						return
+					}
+				} else {
+					http.Error(w, "Cannot start Instance, maximum number of instances reached", http.StatusConflict)
+					return
+				}
+			}
+
+			// If on-demand start is enabled, start the instance
+			if _, err := h.InstanceManager.StartInstance(name); err != nil {
+				http.Error(w, "Failed to start instance: "+err.Error(), http.StatusInternalServerError)
+				return
+			}
+
+			// Wait for the instance to become healthy before proceeding
+			if err := inst.WaitForHealthy(h.cfg.Instances.OnDemandStartTimeout); err != nil { // 2 minutes timeout
+				http.Error(w, "Instance failed to become healthy: "+err.Error(), http.StatusServiceUnavailable)
+				return
+			}
+		}
+
+		proxy, err := inst.GetProxy()
+		if err != nil {
+			http.Error(w, "Failed to get proxy: "+err.Error(), http.StatusInternalServerError)
+			return
+		}
+
+		// Strip the "/llama-cpp/<name>" prefix from the request URL
+		prefix := fmt.Sprintf("/llama-cpp/%s", name)
+		r.URL.Path = strings.TrimPrefix(r.URL.Path, prefix)
+
+		// Update the last request time for the instance
+		inst.UpdateLastRequestTime()
+
+		proxy.ServeHTTP(w, r)
+	}
+}
+
+// ParseCommandRequest represents the request body for command parsing
+type ParseCommandRequest struct {
+	Command string `json:"command"`
+}
+
+// ParseLlamaCommand godoc
+// @Summary Parse llama-server command
+// @Description Parses a llama-server command string into instance options
+// @Tags backends
+// @Security ApiKeyAuth
+// @Accept json
+// @Produce json
+// @Param request body ParseCommandRequest true "Command to parse"
+// @Success 200 {object} instance.CreateInstanceOptions "Parsed options"
+// @Failure 400 {object} map[string]string "Invalid request or command"
+// @Failure 500 {object} map[string]string "Internal Server Error"
+// @Router /backends/llama-cpp/parse-command [post]
+func (h *Handler) ParseLlamaCommand() http.HandlerFunc {
+	type errorResponse struct {
+		Error   string `json:"error"`
+		Details string `json:"details,omitempty"`
+	}
+	writeError := func(w http.ResponseWriter, status int, code, details string) {
+		w.Header().Set("Content-Type", "application/json")
+		w.WriteHeader(status)
+		_ = json.NewEncoder(w).Encode(errorResponse{Error: code, Details: details})
+	}
+	return func(w http.ResponseWriter, r *http.Request) {
+		var req ParseCommandRequest
+		if err := json.NewDecoder(r.Body).Decode(&req); err != nil {
+			writeError(w, http.StatusBadRequest, "invalid_request", "Invalid JSON body")
+			return
+		}
+		if strings.TrimSpace(req.Command) == "" {
+			writeError(w, http.StatusBadRequest, "invalid_command", "Command cannot be empty")
+			return
+		}
+		llamaOptions, err := llamacpp.ParseLlamaCommand(req.Command)
+		if err != nil {
+			writeError(w, http.StatusBadRequest, "parse_error", err.Error())
+			return
+		}
+		options := &instance.CreateInstanceOptions{
+			BackendType:        backends.BackendTypeLlamaCpp,
+			LlamaServerOptions: llamaOptions,
+		}
+		w.Header().Set("Content-Type", "application/json")
+		if err := json.NewEncoder(w).Encode(options); err != nil {
+			writeError(w, http.StatusInternalServerError, "encode_error", err.Error())
+		}
+	}
+}
+
+// ParseMlxCommand godoc
+// @Summary Parse mlx_lm.server command
+// @Description Parses MLX-LM server command string into instance options
+// @Tags backends
+// @Security ApiKeyAuth
+// @Accept json
+// @Produce json
+// @Param request body ParseCommandRequest true "Command to parse"
+// @Success 200 {object} instance.CreateInstanceOptions "Parsed options"
+// @Failure 400 {object} map[string]string "Invalid request or command"
+// @Router /backends/mlx/parse-command [post]
+func (h *Handler) ParseMlxCommand() http.HandlerFunc {
+	type errorResponse struct {
+		Error   string `json:"error"`
+		Details string `json:"details,omitempty"`
+	}
+	writeError := func(w http.ResponseWriter, status int, code, details string) {
+		w.Header().Set("Content-Type", "application/json")
+		w.WriteHeader(status)
+		_ = json.NewEncoder(w).Encode(errorResponse{Error: code, Details: details})
+	}
+	return func(w http.ResponseWriter, r *http.Request) {
+		var req ParseCommandRequest
+		if err := json.NewDecoder(r.Body).Decode(&req); err != nil {
+			writeError(w, http.StatusBadRequest, "invalid_request", "Invalid JSON body")
+			return
+		}
+
+		if strings.TrimSpace(req.Command) == "" {
+			writeError(w, http.StatusBadRequest, "invalid_command", "Command cannot be empty")
+			return
+		}
+
+		mlxOptions, err := mlx.ParseMlxCommand(req.Command)
+		if err != nil {
+			writeError(w, http.StatusBadRequest, "parse_error", err.Error())
+			return
+		}
+
+		// Currently only support mlx_lm backend type
+		backendType := backends.BackendTypeMlxLm
+
+		options := &instance.CreateInstanceOptions{
+			BackendType:      backendType,
+			MlxServerOptions: mlxOptions,
+		}
+
+		w.Header().Set("Content-Type", "application/json")
+		if err := json.NewEncoder(w).Encode(options); err != nil {
+			writeError(w, http.StatusInternalServerError, "encode_error", err.Error())
+		}
+	}
+}
+
+// ParseVllmCommand godoc
+// @Summary Parse vllm serve command
+// @Description Parses a vLLM serve command string into instance options
+// @Tags backends
+// @Security ApiKeyAuth
+// @Accept json
+// @Produce json
+// @Param request body ParseCommandRequest true "Command to parse"
+// @Success 200 {object} instance.CreateInstanceOptions "Parsed options"
+// @Failure 400 {object} map[string]string "Invalid request or command"
+// @Router /backends/vllm/parse-command [post]
+func (h *Handler) ParseVllmCommand() http.HandlerFunc {
+	type errorResponse struct {
+		Error   string `json:"error"`
+		Details string `json:"details,omitempty"`
+	}
+	writeError := func(w http.ResponseWriter, status int, code, details string) {
+		w.Header().Set("Content-Type", "application/json")
+		w.WriteHeader(status)
+		_ = json.NewEncoder(w).Encode(errorResponse{Error: code, Details: details})
+	}
+	return func(w http.ResponseWriter, r *http.Request) {
+		var req ParseCommandRequest
+		if err := json.NewDecoder(r.Body).Decode(&req); err != nil {
+			writeError(w, http.StatusBadRequest, "invalid_request", "Invalid JSON body")
+			return
+		}
+
+		if strings.TrimSpace(req.Command) == "" {
+			writeError(w, http.StatusBadRequest, "invalid_command", "Command cannot be empty")
+			return
+		}
+
+		vllmOptions, err := vllm.ParseVllmCommand(req.Command)
+		if err != nil {
+			writeError(w, http.StatusBadRequest, "parse_error", err.Error())
+			return
+		}
+
+		backendType := backends.BackendTypeVllm
+
+		options := &instance.CreateInstanceOptions{
+			BackendType:       backendType,
+			VllmServerOptions: vllmOptions,
+		}
+
+		w.Header().Set("Content-Type", "application/json")
+		if err := json.NewEncoder(w).Encode(options); err != nil {
+			writeError(w, http.StatusInternalServerError, "encode_error", err.Error())
+		}
+	}
+}

pkg/server/routes.go

@@ -20,7 +20,7 @@ func SetupRouter(handler *Handler) *chi.Mux {
 	r.Use(cors.Handler(cors.Options{
 		AllowedOrigins:   handler.cfg.Server.AllowedOrigins,
 		AllowedMethods:   []string{"GET", "POST", "PUT", "DELETE", "OPTIONS"},
-		AllowedHeaders:   []string{"Accept", "Authorization", "Content-Type", "X-CSRF-Token"},
+		AllowedHeaders:   handler.cfg.Server.AllowedHeaders,
 		ExposedHeaders:   []string{"Link"},
 		AllowCredentials: false,
 		MaxAge:           300,
@@ -44,10 +44,20 @@ func SetupRouter(handler *Handler) *chi.Mux {
 
 		r.Get("/version", handler.VersionHandler()) // Get server version
 
-		r.Route("/server", func(r chi.Router) {
-			r.Get("/help", handler.LlamaServerHelpHandler())
-			r.Get("/version", handler.LlamaServerVersionHandler())
-			r.Get("/devices", handler.LlamaServerListDevicesHandler())
+		// Backend-specific endpoints
+		r.Route("/backends", func(r chi.Router) {
+			r.Route("/llama-cpp", func(r chi.Router) {
+				r.Get("/help", handler.LlamaServerHelpHandler())
+				r.Get("/version", handler.LlamaServerVersionHandler())
+				r.Get("/devices", handler.LlamaServerListDevicesHandler())
+				r.Post("/parse-command", handler.ParseLlamaCommand())
+			})
+			r.Route("/mlx", func(r chi.Router) {
+				r.Post("/parse-command", handler.ParseMlxCommand())
+			})
+			r.Route("/vllm", func(r chi.Router) {
+				r.Post("/parse-command", handler.ParseVllmCommand())
+			})
 		})
 
 		// Instance management endpoints
@@ -93,6 +103,51 @@ func SetupRouter(handler *Handler) *chi.Mux {
 
 	})
 
+	r.Route("/llama-cpp/{name}", func(r chi.Router) {
+
+		// Public Routes
+		// Allow llama-cpp server to serve its own WebUI if it is running.
+		// Don't auto start the server since it can be accessed without an API key
+		r.Get("/", handler.LlamaCppProxy(false))
+
+		// Private Routes
+		r.Group(func(r chi.Router) {
+
+			if authMiddleware != nil && handler.cfg.Auth.RequireInferenceAuth {
+				r.Use(authMiddleware.AuthMiddleware(KeyTypeInference))
+			}
+
+			// This handler auto start the server if it's not running
+			llamaCppHandler := handler.LlamaCppProxy(true)
+
+			// llama.cpp server specific proxy endpoints
+			r.Get("/props", llamaCppHandler)
+			// /slots endpoint is secured (see: https://github.com/ggml-org/llama.cpp/pull/15630)
+			r.Get("/slots", llamaCppHandler)
+			r.Post("/apply-template", llamaCppHandler)
+			r.Post("/completion", llamaCppHandler)
+			r.Post("/detokenize", llamaCppHandler)
+			r.Post("/embeddings", llamaCppHandler)
+			r.Post("/infill", llamaCppHandler)
+			r.Post("/metrics", llamaCppHandler)
+			r.Post("/props", llamaCppHandler)
+			r.Post("/reranking", llamaCppHandler)
+			r.Post("/tokenize", llamaCppHandler)
+
+			// OpenAI-compatible proxy endpoint
+			// Handles all POST requests to /v1/*, including:
+			//   - /v1/completions
+			//   - /v1/chat/completions
+			//   - /v1/embeddings
+			//   - /v1/rerank
+			//   - /v1/reranking
+			// llamaCppHandler is used here because some users of llama.cpp endpoints depend
+			// on "model" field being optional, and handler.OpenAIProxy requires it.
+			r.Post("/v1/*", llamaCppHandler)
+		})
+
+	})
+
 	// Serve WebUI files
 	if err := webui.SetupWebUI(r); err != nil {
 		fmt.Printf("Failed to set up WebUI: %v\n", err)

pkg/validation/validation.go

@@ -2,6 +2,7 @@ package validation
 
 import (
 	"fmt"
+	"llamactl/pkg/backends"
 	"llamactl/pkg/instance"
 	"reflect"
 	"regexp"
@@ -33,20 +34,76 @@ func validateStringForInjection(value string) error {
 	return nil
 }
 
-// ValidateInstanceOptions performs minimal security validation
+// ValidateInstanceOptions performs validation based on backend type
 func ValidateInstanceOptions(options *instance.CreateInstanceOptions) error {
 	if options == nil {
 		return ValidationError(fmt.Errorf("options cannot be nil"))
 	}
 
+	// Validate based on backend type
+	switch options.BackendType {
+	case backends.BackendTypeLlamaCpp:
+		return validateLlamaCppOptions(options)
+	case backends.BackendTypeMlxLm:
+		return validateMlxOptions(options)
+	case backends.BackendTypeVllm:
+		return validateVllmOptions(options)
+	default:
+		return ValidationError(fmt.Errorf("unsupported backend type: %s", options.BackendType))
+	}
+}
+
+// validateLlamaCppOptions validates llama.cpp specific options
+func validateLlamaCppOptions(options *instance.CreateInstanceOptions) error {
+	if options.LlamaServerOptions == nil {
+		return ValidationError(fmt.Errorf("llama server options cannot be nil for llama.cpp backend"))
+	}
+
 	// Use reflection to check all string fields for injection patterns
-	if err := validateStructStrings(&options.LlamaServerOptions, ""); err != nil {
+	if err := validateStructStrings(options.LlamaServerOptions, ""); err != nil {
 		return err
 	}
 
-	// Basic network validation - only check for reasonable ranges
-	if options.Port < 0 || options.Port > 65535 {
-		return ValidationError(fmt.Errorf("invalid port range"))
+	// Basic network validation for port
+	if options.LlamaServerOptions.Port < 0 || options.LlamaServerOptions.Port > 65535 {
+		return ValidationError(fmt.Errorf("invalid port range: %d", options.LlamaServerOptions.Port))
+	}
+
+	return nil
+}
+
+// validateMlxOptions validates MLX backend specific options
+func validateMlxOptions(options *instance.CreateInstanceOptions) error {
+	if options.MlxServerOptions == nil {
+		return ValidationError(fmt.Errorf("MLX server options cannot be nil for MLX backend"))
+	}
+
+	if err := validateStructStrings(options.MlxServerOptions, ""); err != nil {
+		return err
+	}
+
+	// Basic network validation for port
+	if options.MlxServerOptions.Port < 0 || options.MlxServerOptions.Port > 65535 {
+		return ValidationError(fmt.Errorf("invalid port range: %d", options.MlxServerOptions.Port))
+	}
+
+	return nil
+}
+
+// validateVllmOptions validates vLLM backend specific options
+func validateVllmOptions(options *instance.CreateInstanceOptions) error {
+	if options.VllmServerOptions == nil {
+		return ValidationError(fmt.Errorf("vLLM server options cannot be nil for vLLM backend"))
+	}
+
+	// Use reflection to check all string fields for injection patterns
+	if err := validateStructStrings(options.VllmServerOptions, ""); err != nil {
+		return err
+	}
+
+	// Basic network validation for port
+	if options.VllmServerOptions.Port < 0 || options.VllmServerOptions.Port > 65535 {
+		return ValidationError(fmt.Errorf("invalid port range: %d", options.VllmServerOptions.Port))
 	}
 
 	return nil

pkg/validation/validation_test.go

@@ -1,6 +1,7 @@
 package validation_test
 
 import (
+	"llamactl/pkg/backends"
 	"llamactl/pkg/backends/llamacpp"
 	"llamactl/pkg/instance"
 	"llamactl/pkg/testutil"
@@ -83,7 +84,8 @@ func TestValidateInstanceOptions_PortValidation(t *testing.T) {
 	for _, tt := range tests {
 		t.Run(tt.name, func(t *testing.T) {
 			options := &instance.CreateInstanceOptions{
-				LlamaServerOptions: llamacpp.LlamaServerOptions{
+				BackendType: backends.BackendTypeLlamaCpp,
+				LlamaServerOptions: &llamacpp.LlamaServerOptions{
 					Port: tt.port,
 				},
 			}
@@ -136,7 +138,8 @@ func TestValidateInstanceOptions_StringInjection(t *testing.T) {
 		t.Run(tt.name, func(t *testing.T) {
 			// Test with Model field (string field)
 			options := &instance.CreateInstanceOptions{
-				LlamaServerOptions: llamacpp.LlamaServerOptions{
+				BackendType: backends.BackendTypeLlamaCpp,
+				LlamaServerOptions: &llamacpp.LlamaServerOptions{
 					Model: tt.value,
 				},
 			}
@@ -173,7 +176,8 @@ func TestValidateInstanceOptions_ArrayInjection(t *testing.T) {
 		t.Run(tt.name, func(t *testing.T) {
 			// Test with Lora field (array field)
 			options := &instance.CreateInstanceOptions{
-				LlamaServerOptions: llamacpp.LlamaServerOptions{
+				BackendType: backends.BackendTypeLlamaCpp,
+				LlamaServerOptions: &llamacpp.LlamaServerOptions{
 					Lora: tt.array,
 				},
 			}
@@ -196,7 +200,8 @@ func TestValidateInstanceOptions_MultipleFieldInjection(t *testing.T) {
 		{
 			name: "injection in model field",
 			options: &instance.CreateInstanceOptions{
-				LlamaServerOptions: llamacpp.LlamaServerOptions{
+				BackendType: backends.BackendTypeLlamaCpp,
+				LlamaServerOptions: &llamacpp.LlamaServerOptions{
 					Model:  "safe.gguf",
 					HFRepo: "microsoft/model; curl evil.com",
 				},
@@ -206,7 +211,8 @@ func TestValidateInstanceOptions_MultipleFieldInjection(t *testing.T) {
 		{
 			name: "injection in log file",
 			options: &instance.CreateInstanceOptions{
-				LlamaServerOptions: llamacpp.LlamaServerOptions{
+				BackendType: backends.BackendTypeLlamaCpp,
+				LlamaServerOptions: &llamacpp.LlamaServerOptions{
 					Model:   "safe.gguf",
 					LogFile: "/tmp/log.txt | tee /etc/passwd",
 				},
@@ -216,7 +222,8 @@ func TestValidateInstanceOptions_MultipleFieldInjection(t *testing.T) {
 		{
 			name: "all safe fields",
 			options: &instance.CreateInstanceOptions{
-				LlamaServerOptions: llamacpp.LlamaServerOptions{
+				BackendType: backends.BackendTypeLlamaCpp,
+				LlamaServerOptions: &llamacpp.LlamaServerOptions{
 					Model:   "/path/to/model.gguf",
 					HFRepo:  "microsoft/DialoGPT-medium",
 					LogFile: "/tmp/llama.log",
@@ -244,7 +251,8 @@ func TestValidateInstanceOptions_NonStringFields(t *testing.T) {
 		AutoRestart:  testutil.BoolPtr(true),
 		MaxRestarts:  testutil.IntPtr(5),
 		RestartDelay: testutil.IntPtr(10),
-		LlamaServerOptions: llamacpp.LlamaServerOptions{
+		BackendType:  backends.BackendTypeLlamaCpp,
+		LlamaServerOptions: &llamacpp.LlamaServerOptions{
 			Port:        8080,
 			GPULayers:   32,
 			CtxSize:     4096,

webui/package-lock.json

@@ -19,6 +19,7 @@
         "lucide-react": "^0.525.0",
         "react": "^19.1.0",
         "react-dom": "^19.1.0",
+        "sonner": "^2.0.7",
         "tailwind-merge": "^3.3.1",
         "tailwindcss": "^4.1.11",
         "zod": "^4.0.5"
@@ -42,7 +43,7 @@
         "tw-animate-css": "^1.3.5",
         "typescript": "^5.8.3",
         "typescript-eslint": "^8.38.0",
-        "vite": "^7.0.5",
+        "vite": "^7.1.5",
         "vitest": "^3.2.4"
       }
     },
@@ -2109,6 +2110,60 @@
         "node": ">=14.0.0"
       }
     },
+    "node_modules/@tailwindcss/oxide-wasm32-wasi/node_modules/@emnapi/core": {
+      "version": "1.4.3",
+      "inBundle": true,
+      "license": "MIT",
+      "optional": true,
+      "dependencies": {
+        "@emnapi/wasi-threads": "1.0.2",
+        "tslib": "^2.4.0"
+      }
+    },
+    "node_modules/@tailwindcss/oxide-wasm32-wasi/node_modules/@emnapi/runtime": {
+      "version": "1.4.3",
+      "inBundle": true,
+      "license": "MIT",
+      "optional": true,
+      "dependencies": {
+        "tslib": "^2.4.0"
+      }
+    },
+    "node_modules/@tailwindcss/oxide-wasm32-wasi/node_modules/@emnapi/wasi-threads": {
+      "version": "1.0.2",
+      "inBundle": true,
+      "license": "MIT",
+      "optional": true,
+      "dependencies": {
+        "tslib": "^2.4.0"
+      }
+    },
+    "node_modules/@tailwindcss/oxide-wasm32-wasi/node_modules/@napi-rs/wasm-runtime": {
+      "version": "0.2.11",
+      "inBundle": true,
+      "license": "MIT",
+      "optional": true,
+      "dependencies": {
+        "@emnapi/core": "^1.4.3",
+        "@emnapi/runtime": "^1.4.3",
+        "@tybys/wasm-util": "^0.9.0"
+      }
+    },
+    "node_modules/@tailwindcss/oxide-wasm32-wasi/node_modules/@tybys/wasm-util": {
+      "version": "0.9.0",
+      "inBundle": true,
+      "license": "MIT",
+      "optional": true,
+      "dependencies": {
+        "tslib": "^2.4.0"
+      }
+    },
+    "node_modules/@tailwindcss/oxide-wasm32-wasi/node_modules/tslib": {
+      "version": "2.8.0",
+      "inBundle": true,
+      "license": "0BSD",
+      "optional": true
+    },
     "node_modules/@tailwindcss/oxide-win32-arm64-msvc": {
       "version": "4.1.11",
       "resolved": "https://registry.npmjs.org/@tailwindcss/oxide-win32-arm64-msvc/-/oxide-win32-arm64-msvc-4.1.11.tgz",
@@ -4190,10 +4245,13 @@
       }
     },
     "node_modules/fdir": {
-      "version": "6.4.6",
-      "resolved": "https://registry.npmjs.org/fdir/-/fdir-6.4.6.tgz",
-      "integrity": "sha512-hiFoqpyZcfNm1yc4u8oWCf9A2c4D3QjCrks3zmoVKVxpQRzmPNar1hUJcBG2RQHvEVGDN+Jm81ZheVLAQMK6+w==",
+      "version": "6.5.0",
+      "resolved": "https://registry.npmjs.org/fdir/-/fdir-6.5.0.tgz",
+      "integrity": "sha512-tIbYtZbucOs0BRGqPJkshJUYdL+SDH7dVM8gjy+ERp3WAUjLEFJE+02kanyHtwjWOnwrKYBiwAmM0p4kLJAnXg==",
       "license": "MIT",
+      "engines": {
+        "node": ">=12.0.0"
+      },
       "peerDependencies": {
         "picomatch": "^3 || ^4"
       },
@@ -6693,6 +6751,16 @@
         "node": ">=18"
       }
     },
+    "node_modules/sonner": {
+      "version": "2.0.7",
+      "resolved": "https://registry.npmjs.org/sonner/-/sonner-2.0.7.tgz",
+      "integrity": "sha512-W6ZN4p58k8aDKA4XPcx2hpIQXBRAgyiWVkYhT7CvK6D3iAu7xjvVyhQHg2/iaKJZ1XVJ4r7XuwGL+WGEK37i9w==",
+      "license": "MIT",
+      "peerDependencies": {
+        "react": "^18.0.0 || ^19.0.0 || ^19.0.0-rc",
+        "react-dom": "^18.0.0 || ^19.0.0 || ^19.0.0-rc"
+      }
+    },
     "node_modules/source-map-js": {
       "version": "1.2.1",
       "resolved": "https://registry.npmjs.org/source-map-js/-/source-map-js-1.2.1.tgz",
@@ -6973,13 +7041,13 @@
       "license": "MIT"
     },
     "node_modules/tinyglobby": {
-      "version": "0.2.14",
-      "resolved": "https://registry.npmjs.org/tinyglobby/-/tinyglobby-0.2.14.tgz",
-      "integrity": "sha512-tX5e7OM1HnYr2+a2C/4V0htOcSQcoSTH9KgJnVvNm5zm/cyEWKJ7j7YutsH9CxMdtOkkLFy2AHrMci9IM8IPZQ==",
+      "version": "0.2.15",
+      "resolved": "https://registry.npmjs.org/tinyglobby/-/tinyglobby-0.2.15.tgz",
+      "integrity": "sha512-j2Zq4NyQYG5XMST4cbs02Ak8iJUdxRM0XI5QyxXuZOzKOINmWurp3smXu3y5wDcJrptwpSjgXHzIQxR0omXljQ==",
       "license": "MIT",
       "dependencies": {
-        "fdir": "^6.4.4",
-        "picomatch": "^4.0.2"
+        "fdir": "^6.5.0",
+        "picomatch": "^4.0.3"
       },
       "engines": {
         "node": ">=12.0.0"
@@ -7356,17 +7424,17 @@
       }
     },
     "node_modules/vite": {
-      "version": "7.0.5",
-      "resolved": "https://registry.npmjs.org/vite/-/vite-7.0.5.tgz",
-      "integrity": "sha512-1mncVwJxy2C9ThLwz0+2GKZyEXuC3MyWtAAlNftlZZXZDP3AJt5FmwcMit/IGGaNZ8ZOB2BNO/HFUB+CpN0NQw==",
+      "version": "7.1.5",
+      "resolved": "https://registry.npmjs.org/vite/-/vite-7.1.5.tgz",
+      "integrity": "sha512-4cKBO9wR75r0BeIWWWId9XK9Lj6La5X846Zw9dFfzMRw38IlTk2iCcUt6hsyiDRcPidc55ZParFYDXi0nXOeLQ==",
       "license": "MIT",
       "dependencies": {
         "esbuild": "^0.25.0",
-        "fdir": "^6.4.6",
-        "picomatch": "^4.0.2",
+        "fdir": "^6.5.0",
+        "picomatch": "^4.0.3",
         "postcss": "^8.5.6",
-        "rollup": "^4.40.0",
-        "tinyglobby": "^0.2.14"
+        "rollup": "^4.43.0",
+        "tinyglobby": "^0.2.15"
       },
       "bin": {
         "vite": "bin/vite.js"

webui/package.json

@@ -28,6 +28,7 @@
     "lucide-react": "^0.525.0",
     "react": "^19.1.0",
     "react-dom": "^19.1.0",
+    "sonner": "^2.0.7",
     "tailwind-merge": "^3.3.1",
     "tailwindcss": "^4.1.11",
     "zod": "^4.0.5"
@@ -51,7 +52,7 @@
     "tw-animate-css": "^1.3.5",
     "typescript": "^5.8.3",
     "typescript-eslint": "^8.38.0",
-    "vite": "^7.0.5",
+    "vite": "^7.1.5",
     "vitest": "^3.2.4"
   }
 }

webui/src/App.tsx

@@ -8,6 +8,7 @@ import { type CreateInstanceOptions, type Instance } from "@/types/instance";
 import { useInstances } from "@/contexts/InstancesContext";
 import { useAuth } from "@/contexts/AuthContext";
 import { ThemeProvider } from "@/contexts/ThemeContext";
+import { Toaster } from "sonner";
 
 function App() {
   const { isAuthenticated, isLoading: authLoading } = useAuth();
@@ -30,9 +31,9 @@ function App() {
 
   const handleSaveInstance = (name: string, options: CreateInstanceOptions) => {
     if (editingInstance) {
-      updateInstance(editingInstance.name, options);
+      void updateInstance(editingInstance.name, options);
     } else {
-      createInstance(name, options);
+      void createInstance(name, options);
     }
   };
 
@@ -85,6 +86,8 @@ function App() {
           open={isSystemInfoModalOpen}
           onOpenChange={setIsSystemInfoModalOpen}
         />
+        
+        <Toaster />
       </div>
     </ThemeProvider>
   );

webui/src/__tests__/App.test.tsx

@@ -5,6 +5,7 @@ import App from '@/App'
 import { InstancesProvider } from '@/contexts/InstancesContext'
 import { instancesApi } from '@/lib/api'
 import type { Instance } from '@/types/instance'
+import { BackendType } from '@/types/instance'
 import { AuthProvider } from '@/contexts/AuthContext'
 
 // Mock the API
@@ -46,8 +47,8 @@ function renderApp() {
 
 describe('App Component - Critical Business Logic Only', () => {
   const mockInstances: Instance[] = [
-    { name: 'test-instance-1', running: false, options: { model: 'model1.gguf' } },
-    { name: 'test-instance-2', running: true, options: { model: 'model2.gguf' } }
+    { name: 'test-instance-1', status: 'stopped', options: { backend_type: BackendType.LLAMA_CPP, backend_options: { model: 'model1.gguf' } } },
+    { name: 'test-instance-2', status: 'running', options: { backend_type: BackendType.LLAMA_CPP, backend_options: { model: 'model2.gguf' } } }
   ]
 
   beforeEach(() => {
@@ -81,8 +82,8 @@ describe('App Component - Critical Business Logic Only', () => {
       const user = userEvent.setup()
       const newInstance: Instance = {
         name: 'new-test-instance',
-        running: false,
-        options: { model: 'new-model.gguf' }
+        status: 'stopped',
+        options: { backend_type: BackendType.LLAMA_CPP, backend_options: { model: 'new-model.gguf' } }
       }
       vi.mocked(instancesApi.create).mockResolvedValue(newInstance)
 
@@ -105,6 +106,7 @@ describe('App Component - Critical Business Logic Only', () => {
       await waitFor(() => {
         expect(instancesApi.create).toHaveBeenCalledWith('new-test-instance', {
           auto_restart: true, // Default value
+          backend_type: BackendType.LLAMA_CPP
         })
       })
 
@@ -118,8 +120,8 @@ describe('App Component - Critical Business Logic Only', () => {
       const user = userEvent.setup()
       const updatedInstance: Instance = {
         name: 'test-instance-1',
-        running: false,
-        options: { model: 'updated-model.gguf' }
+        status: 'stopped',
+        options: { backend_type: BackendType.LLAMA_CPP, backend_options: { model: 'updated-model.gguf' } }
       }
       vi.mocked(instancesApi.update).mockResolvedValue(updatedInstance)
 
@@ -138,7 +140,8 @@ describe('App Component - Critical Business Logic Only', () => {
       // Verify correct API call with existing instance data
       await waitFor(() => {
         expect(instancesApi.update).toHaveBeenCalledWith('test-instance-1', {
-          model: "model1.gguf", // Pre-filled from existing instance
+          backend_type: BackendType.LLAMA_CPP,
+          backend_options: { model: "model1.gguf" } // Pre-filled from existing instance
         })
       })
     })
@@ -157,7 +160,7 @@ describe('App Component - Critical Business Logic Only', () => {
       expect(screen.getAllByTitle('Start instance').length).toBeGreaterThan(0)
       expect(screen.getAllByTitle('Stop instance').length).toBeGreaterThan(0)
       expect(screen.getAllByTitle('Edit instance').length).toBe(2)
-      expect(screen.getAllByTitle('Delete instance').length).toBeGreaterThan(0)
+      expect(screen.getAllByTitle('More actions').length).toBe(2)
     })
 
     it('delete confirmation calls correct API', async () => {
@@ -171,8 +174,17 @@ describe('App Component - Critical Business Logic Only', () => {
         expect(screen.getByText('test-instance-1')).toBeInTheDocument()
       })
 
-      const deleteButtons = screen.getAllByTitle('Delete instance')
-      await user.click(deleteButtons[0])
+      // First click the "More actions" button to reveal the delete button
+      const moreActionsButtons = screen.getAllByTitle('More actions')
+      await user.click(moreActionsButtons[0])
+
+      // Wait for the delete button to appear and click it
+      await waitFor(() => {
+        expect(screen.getByTitle('Delete instance')).toBeInTheDocument()
+      })
+
+      const deleteButton = screen.getByTitle('Delete instance')
+      await user.click(deleteButton)
 
       // Verify confirmation and API call
       expect(confirmSpy).toHaveBeenCalledWith('Are you sure you want to delete instance "test-instance-1"?')

webui/src/components/BackendBadge.tsx

@@ -0,0 +1,65 @@
+import React from "react";
+import { Badge } from "@/components/ui/badge";
+import { BackendType, type BackendTypeValue } from "@/types/instance";
+import { Server, Package } from "lucide-react";
+
+interface BackendBadgeProps {
+  backend?: BackendTypeValue;
+  docker?: boolean;
+}
+
+const BackendBadge: React.FC<BackendBadgeProps> = ({ backend, docker }) => {
+  if (!backend) {
+    return null;
+  }
+
+  const getText = () => {
+    switch (backend) {
+      case BackendType.LLAMA_CPP:
+        return "llama.cpp";
+      case BackendType.MLX_LM:
+        return "MLX";
+      case BackendType.VLLM:
+        return "vLLM";
+      default:
+        return backend;
+    }
+  };
+
+  const getColorClasses = () => {
+    switch (backend) {
+      case BackendType.LLAMA_CPP:
+        return "bg-blue-100 text-blue-800 border-blue-200 dark:bg-blue-900 dark:text-blue-200 dark:border-blue-800";
+      case BackendType.MLX_LM:
+        return "bg-green-100 text-green-800 border-green-200 dark:bg-green-900 dark:text-green-200 dark:border-green-800";
+      case BackendType.VLLM:
+        return "bg-purple-100 text-purple-800 border-purple-200 dark:bg-purple-900 dark:text-purple-200 dark:border-purple-800";
+      default:
+        return "bg-gray-100 text-gray-800 border-gray-200 dark:bg-gray-900 dark:text-gray-200 dark:border-gray-800";
+    }
+  };
+
+  return (
+    <div className="flex items-center gap-1">
+      <Badge
+        variant="outline"
+        className={`flex items-center gap-1.5 ${getColorClasses()}`}
+      >
+        <Server className="h-3 w-3" />
+        <span className="text-xs">{getText()}</span>
+      </Badge>
+      {docker && (
+        <Badge
+          variant="outline"
+          className="flex items-center gap-1.5 bg-orange-100 text-orange-800 border-orange-200 dark:bg-orange-900 dark:text-orange-200 dark:border-orange-800"
+          title="Docker enabled"
+        >
+          <Package className="h-3 w-3" />
+          <span className="text-[10px] uppercase tracking-wide">Docker</span>
+        </Badge>
+      )}
+    </div>
+  );
+};
+
+export default BackendBadge;

webui/src/components/BackendFormField.tsx

@@ -2,24 +2,23 @@ import React from 'react'
 import { Input } from '@/components/ui/input'
 import { Label } from '@/components/ui/label'
 import { Checkbox } from '@/components/ui/checkbox'
-import type { CreateInstanceOptions } from '@/types/instance'
-import { getFieldType, basicFieldsConfig } from '@/lib/zodFormUtils'
+import { getBackendFieldType, basicBackendFieldsConfig } from '@/lib/zodFormUtils'
 
-interface ZodFormFieldProps {
-  fieldKey: keyof CreateInstanceOptions
+interface BackendFormFieldProps {
+  fieldKey: string
   value: string | number | boolean | string[] | undefined
-  onChange: (key: keyof CreateInstanceOptions, value: string | number | boolean | string[] | undefined) => void
+  onChange: (key: string, value: string | number | boolean | string[] | undefined) => void
 }
 
-const ZodFormField: React.FC<ZodFormFieldProps> = ({ fieldKey, value, onChange }) => {
+const BackendFormField: React.FC<BackendFormFieldProps> = ({ fieldKey, value, onChange }) => {
   // Get configuration for basic fields, or use field name for advanced fields
-  const config = basicFieldsConfig[fieldKey as string] || { label: fieldKey }
+  const config = basicBackendFieldsConfig[fieldKey as string] || { label: fieldKey }
   
   // Get type from Zod schema
-  const fieldType = getFieldType(fieldKey)
+  const fieldType = getBackendFieldType(fieldKey)
 
   const handleChange = (newValue: string | number | boolean | string[] | undefined) => {
-    onChange(fieldKey, newValue)
+    onChange(fieldKey as string, newValue)
   }
 
   const renderField = () => {
@@ -46,7 +45,6 @@ const ZodFormField: React.FC<ZodFormFieldProps> = ({ fieldKey, value, onChange }
           <div className="grid gap-2">
             <Label htmlFor={fieldKey}>
               {config.label}
-              {config.required && <span className="text-red-500 ml-1">*</span>}
             </Label>
             <Input
               id={fieldKey}
@@ -73,7 +71,6 @@ const ZodFormField: React.FC<ZodFormFieldProps> = ({ fieldKey, value, onChange }
           <div className="grid gap-2">
             <Label htmlFor={fieldKey}>
               {config.label}
-              {config.required && <span className="text-red-500 ml-1">*</span>}
             </Label>
             <Input
               id={fieldKey}
@@ -100,7 +97,6 @@ const ZodFormField: React.FC<ZodFormFieldProps> = ({ fieldKey, value, onChange }
           <div className="grid gap-2">
             <Label htmlFor={fieldKey}>
               {config.label}
-              {config.required && <span className="text-red-500 ml-1">*</span>}
             </Label>
             <Input
               id={fieldKey}
@@ -120,4 +116,4 @@ const ZodFormField: React.FC<ZodFormFieldProps> = ({ fieldKey, value, onChange }
   return <div className="space-y-2">{renderField()}</div>
 }
 
-export default ZodFormField
+export default BackendFormField

webui/src/components/HealthBadge.tsx

@@ -27,6 +27,8 @@ const HealthBadge: React.FC<HealthBadgeProps> = ({ health }) => {
         return <XCircle className="h-3 w-3" />;
       case "unknown":
         return <Loader2 className="h-3 w-3 animate-spin" />;
+      case "failed":
+        return <XCircle className="h-3 w-3" />;
     }
   };
 
@@ -40,6 +42,8 @@ const HealthBadge: React.FC<HealthBadgeProps> = ({ health }) => {
         return "destructive";
       case "unknown":
         return "secondary";
+      case "failed":
+        return "destructive";
     }
   };
 
@@ -53,6 +57,8 @@ const HealthBadge: React.FC<HealthBadgeProps> = ({ health }) => {
         return "Error";
       case "unknown":
         return "Unknown";
+      case "failed":
+        return "Failed";
     }
   };
 

webui/src/components/InstanceCard.tsx

@@ -2,9 +2,10 @@
 import { Button } from "@/components/ui/button";
 import { Card, CardContent, CardHeader, CardTitle } from "@/components/ui/card";
 import type { Instance } from "@/types/instance";
-import { Edit, FileText, Play, Square, Trash2 } from "lucide-react";
+import { Edit, FileText, Play, Square, Trash2, MoreHorizontal } from "lucide-react";
 import LogsDialog from "@/components/LogDialog";
 import HealthBadge from "@/components/HealthBadge";
+import BackendBadge from "@/components/BackendBadge";
 import { useState } from "react";
 import { useInstanceHealth } from "@/hooks/useInstanceHealth";
 
@@ -24,7 +25,8 @@ function InstanceCard({
   editInstance,
 }: InstanceCardProps) {
   const [isLogsOpen, setIsLogsOpen] = useState(false);
-  const health = useInstanceHealth(instance.name, instance.running);
+  const [showAllActions, setShowAllActions] = useState(false);
+  const health = useInstanceHealth(instance.name, instance.status);
 
   const handleStart = () => {
     startInstance(instance.name);
@@ -50,38 +52,48 @@ function InstanceCard({
     setIsLogsOpen(true);
   };
 
+  const running = instance.status === "running";
+
   return (
     <>
-      <Card>
-        <CardHeader className="pb-3">
-          <div className="flex items-center justify-between">
-            <CardTitle className="text-lg">{instance.name}</CardTitle>
-            {instance.running && <HealthBadge health={health} />}
+      <Card className="hover:shadow-md transition-shadow">
+        <CardHeader className="pb-4">
+          {/* Header with instance name and status badges */}
+          <div className="space-y-3">
+            <CardTitle className="text-lg font-semibold leading-tight break-words">
+              {instance.name}
+            </CardTitle>
+            
+            {/* Badges row */}
+            <div className="flex items-center gap-2 flex-wrap">
+              <BackendBadge backend={instance.options?.backend_type} docker={instance.docker_enabled} />
+              {running && <HealthBadge health={health} />}
+            </div>
           </div>
         </CardHeader>
 
-        <CardContent>
-          <div className="flex gap-1">
+        <CardContent className="pt-0">
+          {/* Primary actions - always visible */}
+          <div className="flex items-center gap-2 mb-3">
             <Button
               size="sm"
-              variant="outline"
-              onClick={handleStart}
-              disabled={instance.running}
-              title="Start instance"
-              data-testid="start-instance-button"
+              variant={running ? "outline" : "default"}
+              onClick={running ? handleStop : handleStart}
+              className="flex-1"
+              title={running ? "Stop instance" : "Start instance"}
+              data-testid={running ? "stop-instance-button" : "start-instance-button"}
             >
-              <Play className="h-4 w-4" />
-            </Button>
-
-            <Button
-              size="sm"
-              variant="outline"
-              onClick={handleStop}
-              disabled={!instance.running}
-              title="Stop instance"
-              data-testid="stop-instance-button"
-            >
-              <Square className="h-4 w-4" />
+              {running ? (
+                <>
+                  <Square className="h-4 w-4 mr-1" />
+                  Stop
+                </>
+              ) : (
+                <>
+                  <Play className="h-4 w-4 mr-1" />
+                  Start
+                </>
+              )}
             </Button>
 
             <Button
@@ -97,24 +109,40 @@ function InstanceCard({
             <Button
               size="sm"
               variant="outline"
-              onClick={handleLogs}
-              title="View logs"
-              data-testid="view-logs-button"
+              onClick={() => setShowAllActions(!showAllActions)}
+              title="More actions"
             >
-              <FileText className="h-4 w-4" />
-            </Button>
-
-            <Button
-              size="sm"
-              variant="destructive"
-              onClick={handleDelete}
-              disabled={instance.running}
-              title="Delete instance"
-              data-testid="delete-instance-button"
-            >
-              <Trash2 className="h-4 w-4" />
+              <MoreHorizontal className="h-4 w-4" />
             </Button>
           </div>
+
+          {/* Secondary actions - collapsible */}
+          {showAllActions && (
+            <div className="flex items-center gap-2 pt-2 border-t border-border">
+              <Button
+                size="sm"
+                variant="outline"
+                onClick={handleLogs}
+                title="View logs"
+                data-testid="view-logs-button"
+                className="flex-1"
+              >
+                <FileText className="h-4 w-4 mr-1" />
+                Logs
+              </Button>
+
+              <Button
+                size="sm"
+                variant="destructive"
+                onClick={handleDelete}
+                disabled={running}
+                title="Delete instance"
+                data-testid="delete-instance-button"
+              >
+                <Trash2 className="h-4 w-4" />
+              </Button>
+            </div>
+          )}
         </CardContent>
       </Card>
 
@@ -122,7 +150,7 @@ function InstanceCard({
         open={isLogsOpen}
         onOpenChange={setIsLogsOpen}
         instanceName={instance.name}
-        isRunning={instance.running}
+        isRunning={running}
       />
     </>
   );

webui/src/components/InstanceDialog.tsx

@@ -1,7 +1,5 @@
 import React, { useState, useEffect } from "react";
 import { Button } from "@/components/ui/button";
-import { Input } from "@/components/ui/input";
-import { Label } from "@/components/ui/label";
 import {
   Dialog,
   DialogContent,
@@ -10,10 +8,10 @@ import {
   DialogHeader,
   DialogTitle,
 } from "@/components/ui/dialog";
-import type { CreateInstanceOptions, Instance } from "@/types/instance";
-import { getBasicFields, getAdvancedFields } from "@/lib/zodFormUtils";
-import { ChevronDown, ChevronRight } from "lucide-react";
-import ZodFormField from "@/components/ZodFormField";
+import { BackendType, type CreateInstanceOptions, type Instance } from "@/types/instance";
+import ParseCommandDialog from "@/components/ParseCommandDialog";
+import InstanceSettingsCard from "@/components/instance/InstanceSettingsCard";
+import BackendConfigurationCard from "@/components/instance/BackendConfigurationCard";
 
 interface InstanceDialogProps {
   open: boolean;
@@ -29,16 +27,12 @@ const InstanceDialog: React.FC<InstanceDialogProps> = ({
   instance,
 }) => {
   const isEditing = !!instance;
-  const isRunning = instance?.running || true; // Assume running if instance exists
 
   const [instanceName, setInstanceName] = useState("");
   const [formData, setFormData] = useState<CreateInstanceOptions>({});
-  const [showAdvanced, setShowAdvanced] = useState(false);
   const [nameError, setNameError] = useState("");
+  const [showParseDialog, setShowParseDialog] = useState(false);
 
-  // Get field lists dynamically from the type
-  const basicFields = getBasicFields();
-  const advancedFields = getAdvancedFields();
 
   // Reset form when dialog opens/closes or when instance changes
   useEffect(() => {
@@ -52,17 +46,39 @@ const InstanceDialog: React.FC<InstanceDialogProps> = ({
         setInstanceName("");
         setFormData({
           auto_restart: true, // Default value
+          backend_type: BackendType.LLAMA_CPP, // Default backend type
+          backend_options: {},
         });
       }
-      setShowAdvanced(false); // Always start with basic view
       setNameError(""); // Reset any name errors
     }
   }, [open, instance]);
 
   const handleFieldChange = (key: keyof CreateInstanceOptions, value: any) => {
+    setFormData((prev) => {
+      // If backend_type is changing, clear backend_options
+      if (key === 'backend_type' && prev.backend_type !== value) {
+        return {
+          ...prev,
+          [key]: value,
+          backend_options: {}, // Clear backend options when backend type changes
+        };
+      }
+      
+      return {
+        ...prev,
+        [key]: value,
+      };
+    });
+  };
+
+  const handleBackendFieldChange = (key: string, value: any) => {
     setFormData((prev) => ({
       ...prev,
-      [key]: value,
+      backend_options: {
+        ...prev.backend_options,
+        [key]: value,
+      } as any,
     }));
   };
 
@@ -90,7 +106,24 @@ const InstanceDialog: React.FC<InstanceDialogProps> = ({
     // Clean up undefined values to avoid sending empty fields
     const cleanOptions: CreateInstanceOptions = {};
     Object.entries(formData).forEach(([key, value]) => {
-      if (value !== undefined && value !== "" && value !== null) {
+      if (key === 'backend_options' && value && typeof value === 'object') {
+        // Handle backend_options specially - clean nested object
+        const cleanBackendOptions: any = {};
+        Object.entries(value).forEach(([backendKey, backendValue]) => {
+          if (backendValue !== undefined && backendValue !== null && (typeof backendValue !== 'string' || backendValue.trim() !== "")) {
+            // Handle arrays - don't include empty arrays
+            if (Array.isArray(backendValue) && backendValue.length === 0) {
+              return;
+            }
+            cleanBackendOptions[backendKey] = backendValue;
+          }
+        });
+        
+        // Only include backend_options if it has content
+        if (Object.keys(cleanBackendOptions).length > 0) {
+          (cleanOptions as any)[key] = cleanBackendOptions;
+        }
+      } else if (value !== undefined && value !== null && (typeof value !== 'string' || value.trim() !== "")) {
         // Handle arrays - don't include empty arrays
         if (Array.isArray(value) && value.length === 0) {
           return;
@@ -107,12 +140,25 @@ const InstanceDialog: React.FC<InstanceDialogProps> = ({
     onOpenChange(false);
   };
 
-  const toggleAdvanced = () => {
-    setShowAdvanced(!showAdvanced);
+
+  const handleCommandParsed = (parsedOptions: CreateInstanceOptions) => {
+    setFormData(prev => ({
+      ...prev,
+      ...parsedOptions,
+    }));
+    setShowParseDialog(false);
   };
 
-  // Check if auto_restart is enabled
-  const isAutoRestartEnabled = formData.auto_restart === true;
+
+  // Save button label logic
+  let saveButtonLabel = "Create Instance";
+  if (isEditing) {
+    if (instance?.status === "running") {
+      saveButtonLabel = "Update & Restart Instance";
+    } else {
+      saveButtonLabel = "Update Instance";
+    }
+  }
 
   return (
     <Dialog open={open} onOpenChange={onOpenChange}>
@@ -129,125 +175,25 @@ const InstanceDialog: React.FC<InstanceDialogProps> = ({
         </DialogHeader>
 
         <div className="flex-1 overflow-y-auto">
-          <div className="grid gap-6 py-4">
-            {/* Instance Name - Special handling since it's not in CreateInstanceOptions */}
-            <div className="grid gap-2">
-              <Label htmlFor="name">
-                Instance Name <span className="text-red-500">*</span>
-              </Label>
-              <Input
-                id="name"
-                value={instanceName}
-                onChange={(e) => handleNameChange(e.target.value)}
-                placeholder="my-instance"
-                disabled={isEditing} // Don't allow name changes when editing
-                className={nameError ? "border-red-500" : ""}
-              />
-              {nameError && <p className="text-sm text-red-500">{nameError}</p>}
-              <p className="text-sm text-muted-foreground">
-                Unique identifier for the instance
-              </p>
-            </div>
+          <div className="space-y-6 py-4">
+            {/* Instance Settings Card */}
+            <InstanceSettingsCard
+              instanceName={instanceName}
+              nameError={nameError}
+              isEditing={isEditing}
+              formData={formData}
+              onNameChange={handleNameChange}
+              onChange={handleFieldChange}
+            />
 
-            {/* Auto Restart Configuration Section */}
-            <div className="space-y-4">
-              <h3 className="text-lg font-medium">
-                Auto Restart Configuration
-              </h3>
+            {/* Backend Configuration Card */}
+            <BackendConfigurationCard
+              formData={formData}
+              onBackendFieldChange={handleBackendFieldChange}
+              onChange={handleFieldChange}
+              onParseCommand={() => setShowParseDialog(true)}
+            />
 
-              {/* Auto Restart Toggle */}
-              <ZodFormField
-                fieldKey="auto_restart"
-                value={formData.auto_restart}
-                onChange={handleFieldChange}
-              />
-
-              {/* Show restart options only when auto restart is enabled */}
-              {isAutoRestartEnabled && (
-                <div className="ml-6 space-y-4 border-l-2 border-muted pl-4">
-                  <ZodFormField
-                    fieldKey="max_restarts"
-                    value={formData.max_restarts}
-                    onChange={handleFieldChange}
-                  />
-                  <ZodFormField
-                    fieldKey="restart_delay"
-                    value={formData.restart_delay}
-                    onChange={handleFieldChange}
-                  />
-                </div>
-              )}
-            </div>
-
-            {/* Basic Fields - Automatically generated from type (excluding auto restart options) */}
-            <div className="space-y-4">
-              <h3 className="text-lg font-medium">Basic Configuration</h3>
-              {basicFields
-                .filter(
-                  (fieldKey) =>
-                    fieldKey !== "auto_restart" &&
-                    fieldKey !== "max_restarts" &&
-                    fieldKey !== "restart_delay"
-                ) // Exclude auto_restart, max_restarts, and restart_delay as they're handled above
-                .map((fieldKey) => (
-                  <ZodFormField
-                    key={fieldKey}
-                    fieldKey={fieldKey}
-                    value={formData[fieldKey]}
-                    onChange={handleFieldChange}
-                  />
-                ))}
-            </div>
-
-            {/* Advanced Fields Toggle */}
-            <div className="border-t pt-4">
-              <Button
-                variant="ghost"
-                onClick={toggleAdvanced}
-                className="flex items-center gap-2 p-0 h-auto font-medium"
-              >
-                {showAdvanced ? (
-                  <ChevronDown className="h-4 w-4" />
-                ) : (
-                  <ChevronRight className="h-4 w-4" />
-                )}
-                Advanced Configuration
-                <span className="text-muted-foreground text-sm font-normal">
-                  (
-                  {
-                    advancedFields.filter(
-                      (f) =>
-                        !["max_restarts", "restart_delay"].includes(f as string)
-                    ).length
-                  }{" "}
-                  options)
-                </span>
-              </Button>
-            </div>
-
-            {/* Advanced Fields - Automatically generated from type (excluding restart options) */}
-            {showAdvanced && (
-              <div className="space-y-4 pl-6 border-l-2 border-muted">
-                <div className="space-y-4">
-                  {advancedFields
-                    .filter(
-                      (fieldKey) =>
-                        !["max_restarts", "restart_delay"].includes(
-                          fieldKey as string
-                        )
-                    ) // Exclude restart options as they're handled above
-                    .sort()
-                    .map((fieldKey) => (
-                      <ZodFormField
-                        key={fieldKey}
-                        fieldKey={fieldKey}
-                        value={formData[fieldKey]}
-                        onChange={handleFieldChange}
-                      />
-                    ))}
-                </div>
-              </div>
-            )}
           </div>
         </div>
 
@@ -264,14 +210,17 @@ const InstanceDialog: React.FC<InstanceDialogProps> = ({
             disabled={!instanceName.trim() || !!nameError}
             data-testid="dialog-save-button"
           >
-            {isEditing
-              ? isRunning
-                ? "Update & Restart Instance"
-                : "Update Instance"
-              : "Create Instance"}
+            {saveButtonLabel}
           </Button>
         </DialogFooter>
       </DialogContent>
+      
+      <ParseCommandDialog
+        open={showParseDialog}
+        onOpenChange={setShowParseDialog}
+        onParsed={handleCommandParsed}
+        backendType={formData.backend_type || BackendType.LLAMA_CPP}
+      />
     </Dialog>
   );
 };

webui/src/components/ParseCommandDialog.tsx

@@ -0,0 +1,151 @@
+import React, { useState } from "react";
+import { Button } from "@/components/ui/button";
+import { Label } from "@/components/ui/label";
+import {
+  Dialog,
+  DialogContent,
+  DialogDescription,
+  DialogFooter,
+  DialogHeader,
+  DialogTitle,
+} from "@/components/ui/dialog";
+import { BackendType, type BackendTypeValue, type CreateInstanceOptions } from "@/types/instance";
+import { backendsApi } from "@/lib/api";
+import { toast } from "sonner";
+
+interface ParseCommandDialogProps {
+  open: boolean;
+  onOpenChange: (open: boolean) => void;
+  onParsed: (options: CreateInstanceOptions) => void;
+  backendType: BackendTypeValue;
+}
+
+const ParseCommandDialog: React.FC<ParseCommandDialogProps> = ({
+  open,
+  onOpenChange,
+  onParsed,
+  backendType,
+}) => {
+  const [command, setCommand] = useState('');
+  const [loading, setLoading] = useState(false);
+  const [error, setError] = useState<string | null>(null);
+
+  const handleParse = async () => {
+    if (!command.trim()) {
+      setError("Command cannot be empty");
+      return;
+    }
+
+    setLoading(true);
+    setError(null);
+
+    try {
+      let options: CreateInstanceOptions;
+
+      // Parse based on selected backend type
+      switch (backendType) {
+        case BackendType.LLAMA_CPP:
+          options = await backendsApi.llamaCpp.parseCommand(command);
+          break;
+        case BackendType.MLX_LM:
+          options = await backendsApi.mlx.parseCommand(command);
+          break;
+        case BackendType.VLLM:
+          options = await backendsApi.vllm.parseCommand(command);
+          break;
+        default:
+          throw new Error(`Unsupported backend type: ${backendType}`);
+      }
+
+      onParsed(options);
+      onOpenChange(false);
+      setCommand('');
+      setError(null);
+      toast.success('Command parsed successfully');
+    } catch (err) {
+      const errorMessage = err instanceof Error ? err.message : 'Failed to parse command';
+      setError(errorMessage);
+      toast.error('Failed to parse command', {
+        description: errorMessage
+      });
+    } finally {
+      setLoading(false);
+    }
+  };
+
+  const handleOpenChange = (open: boolean) => {
+    if (!open) {
+      setCommand('');
+      setError(null);
+    }
+    onOpenChange(open);
+  };
+
+  const backendPlaceholders: Record<BackendTypeValue, string> = {
+    [BackendType.LLAMA_CPP]: "llama-server --model /path/to/model.gguf --gpu-layers 32 --ctx-size 4096",
+    [BackendType.MLX_LM]: "mlx_lm.server --model mlx-community/Mistral-7B-Instruct-v0.3-4bit --host 0.0.0.0 --port 8080",
+    [BackendType.VLLM]: "vllm serve microsoft/DialoGPT-medium --tensor-parallel-size 2 --gpu-memory-utilization 0.9",
+  };
+
+  const getPlaceholderForBackend = (backendType: BackendTypeValue): string => {
+    return backendPlaceholders[backendType] || "Enter your command here...";
+  };
+
+  return (
+    <Dialog open={open} onOpenChange={handleOpenChange}>
+      <DialogContent className="sm:max-w-[600px]">
+        <DialogHeader>
+          <DialogTitle>Parse Backend Command</DialogTitle>
+          <DialogDescription>
+            Select your backend type and paste the command to automatically populate the form fields
+          </DialogDescription>
+        </DialogHeader>
+
+        <div className="space-y-4">
+          <div>
+            <Label className="text-sm font-medium">Backend Type:
+              <span className="font-normal text-muted-foreground">
+                {backendType === BackendType.LLAMA_CPP && 'Llama Server'}
+                {backendType === BackendType.MLX_LM && 'MLX LM'}
+                {backendType === BackendType.VLLM && 'vLLM'}
+              </span>
+            </Label>
+          </div>
+
+          <div>
+            <Label htmlFor="command">Command</Label>
+            <textarea
+              id="command"
+              value={command}
+              onChange={(e) => setCommand(e.target.value)}
+              placeholder={getPlaceholderForBackend(backendType)}
+              className="w-full h-32 p-3 mt-2 border border-input rounded-md font-mono text-sm resize-vertical focus:outline-none focus:ring-2 focus:ring-ring focus:ring-offset-2"
+            />
+          </div>
+
+          {error && (
+            <div className="text-destructive text-sm bg-destructive/10 p-3 rounded-md">
+              {error}
+            </div>
+          )}
+        </div>
+
+        <DialogFooter>
+          <Button variant="outline" onClick={() => handleOpenChange(false)}>
+            Cancel
+          </Button>
+          <Button 
+            onClick={() => {
+              handleParse().catch(console.error);
+            }}
+            disabled={!command.trim() || loading}
+          >
+            {loading ? 'Parsing...' : 'Parse Command'}
+          </Button>
+        </DialogFooter>
+      </DialogContent>
+    </Dialog>
+  );
+};
+
+export default ParseCommandDialog;

webui/src/components/SystemInfoDialog.tsx

@@ -8,16 +8,19 @@ import {
   DialogHeader,
   DialogTitle,
 } from '@/components/ui/dialog'
-import { 
-  RefreshCw, 
+import SelectInput from '@/components/form/SelectInput'
+import {
+  RefreshCw,
   AlertCircle,
   Loader2,
   ChevronDown,
   ChevronRight,
   Monitor,
-  HelpCircle
+  HelpCircle,
+  Info
 } from 'lucide-react'
 import { serverApi } from '@/lib/api'
+import { BackendType, type BackendTypeValue } from '@/types/instance'
 
 // Helper to get version from environment
 const getAppVersion = (): string => {
@@ -28,166 +31,234 @@ const getAppVersion = (): string => {
   }
 }
 
-interface SystemInfoModalProps {
+interface SystemInfoDialogProps {
   open: boolean
   onOpenChange: (open: boolean) => void
 }
 
-interface SystemInfo {
+interface BackendInfo {
   version: string
   devices: string
   help: string
 }
 
-const SystemInfoDialog: React.FC<SystemInfoModalProps> = ({
+const BACKEND_OPTIONS = [
+  { value: BackendType.LLAMA_CPP, label: 'Llama Server' },
+  { value: BackendType.MLX_LM, label: 'MLX LM' },
+  { value: BackendType.VLLM, label: 'vLLM' },
+]
+
+const SystemInfoDialog: React.FC<SystemInfoDialogProps> = ({
   open,
   onOpenChange
 }) => {
-  const [systemInfo, setSystemInfo] = useState<SystemInfo | null>(null)
+  const [selectedBackend, setSelectedBackend] = useState<BackendTypeValue>(BackendType.LLAMA_CPP)
+  const [backendInfo, setBackendInfo] = useState<BackendInfo | null>(null)
   const [loading, setLoading] = useState(false)
   const [error, setError] = useState<string | null>(null)
   const [showHelp, setShowHelp] = useState(false)
 
-  // Fetch system info
-  const fetchSystemInfo = async () => {
+  // Fetch backend info
+  const fetchBackendInfo = async (backend: BackendTypeValue) => {
+    if (backend !== BackendType.LLAMA_CPP) {
+      setBackendInfo(null)
+      setError(null)
+      return
+    }
+
     setLoading(true)
     setError(null)
-    
+
     try {
       const [version, devices, help] = await Promise.all([
         serverApi.getVersion(),
         serverApi.getDevices(),
         serverApi.getHelp()
       ])
-      
-      setSystemInfo({ version, devices, help })
+
+      setBackendInfo({ version, devices, help })
     } catch (err) {
-      setError(err instanceof Error ? err.message : 'Failed to fetch system info')
+      setError(err instanceof Error ? err.message : 'Failed to fetch backend info')
     } finally {
       setLoading(false)
     }
   }
 
-  // Load data when dialog opens
+  // Load data when dialog opens or backend changes
   useEffect(() => {
     if (open) {
-      fetchSystemInfo()
+      void fetchBackendInfo(selectedBackend)
     }
-  }, [open])
+  }, [open, selectedBackend])
+
+  const handleBackendChange = (value: string) => {
+    setSelectedBackend(value as BackendTypeValue)
+    setShowHelp(false) // Reset help section when switching backends
+  }
+
+  const renderBackendSpecificContent = () => {
+    if (selectedBackend !== BackendType.LLAMA_CPP) {
+      return (
+        <div className="flex items-center justify-center py-8">
+          <div className="text-center space-y-3">
+            <Info className="h-8 w-8 text-gray-400 mx-auto" />
+            <div>
+              <h3 className="font-semibold text-gray-700">Backend Info Not Available</h3>
+              <p className="text-sm text-gray-500 mt-1">
+                Information for {BACKEND_OPTIONS.find(b => b.value === selectedBackend)?.label} backend is not yet implemented.
+              </p>
+            </div>
+          </div>
+        </div>
+      )
+    }
+
+    if (loading && !backendInfo) {
+      return (
+        <div className="flex items-center justify-center py-8">
+          <Loader2 className="h-6 w-6 animate-spin text-gray-400" />
+          <span className="ml-2 text-gray-400">Loading backend information...</span>
+        </div>
+      )
+    }
+
+    if (error) {
+      return (
+        <div className="flex items-center gap-2 p-4 bg-destructive/10 border border-destructive/20 rounded-lg">
+          <AlertCircle className="h-4 w-4 text-destructive" />
+          <span className="text-sm text-destructive">{error}</span>
+        </div>
+      )
+    }
+
+    if (!backendInfo) {
+      return null
+    }
+
+    return (
+      <div className="space-y-6">
+        {/* Backend Version Section */}
+        <div className="space-y-3">
+          <h3 className="font-semibold">
+            {BACKEND_OPTIONS.find(b => b.value === selectedBackend)?.label} Version
+          </h3>
+
+          <div className="bg-gray-900 rounded-lg p-4">
+            <div className="mb-2">
+              <span className="text-sm text-gray-400">$ llama-server --version</span>
+            </div>
+            <pre className="text-sm text-gray-300 whitespace-pre-wrap font-mono">
+              {backendInfo.version}
+            </pre>
+          </div>
+        </div>
+
+        {/* Devices Section */}
+        <div className="space-y-3">
+          <div className="flex items-center gap-2">
+            <h3 className="font-semibold">Available Devices</h3>
+          </div>
+
+          <div className="bg-gray-900 rounded-lg p-4">
+            <div className="mb-2">
+              <span className="text-sm text-gray-400">$ llama-server --list-devices</span>
+            </div>
+            <pre className="text-sm text-gray-300 whitespace-pre-wrap font-mono">
+              {backendInfo.devices}
+            </pre>
+          </div>
+        </div>
+
+        {/* Help Section */}
+        <div className="space-y-3">
+          <Button
+            variant="ghost"
+            onClick={() => setShowHelp(!showHelp)}
+            className="flex items-center gap-2 p-0 h-auto font-semibold"
+          >
+            {showHelp ? (
+              <ChevronDown className="h-4 w-4" />
+            ) : (
+              <ChevronRight className="h-4 w-4" />
+            )}
+            <HelpCircle className="h-4 w-4" />
+            Command Line Options
+          </Button>
+
+          {showHelp && (
+            <div className="bg-gray-900 rounded-lg p-4">
+              <div className="mb-2">
+                <span className="text-sm text-gray-400">$ llama-server --help</span>
+              </div>
+              <pre className="text-sm text-gray-300 whitespace-pre-wrap font-mono max-h-64 overflow-y-auto">
+                {backendInfo.help}
+              </pre>
+            </div>
+          )}
+        </div>
+      </div>
+    )
+  }
 
   return (
-    <Dialog open={open} onOpenChange={onOpenChange} >
+    <Dialog open={open} onOpenChange={onOpenChange}>
       <DialogContent className="sm:max-w-4xl max-w-[calc(100%-2rem)] max-h-[80vh] flex flex-col">
         <DialogHeader>
-          <div className="flex items-center justify-between">
-            <div>
-              <DialogTitle className="flex items-center gap-2">
-                <Monitor className="h-5 w-5" />
-                System Information
-              </DialogTitle>
-              <DialogDescription>
-                Llama.cpp server environment and capabilities
-              </DialogDescription>
-            </div>
-            
-            <Button
-              variant="outline"
-              size="sm"
-              onClick={fetchSystemInfo}
-              disabled={loading}
-            >
-              {loading ? (
-                <Loader2 className="h-4 w-4 animate-spin" />
-              ) : (
-                <RefreshCw className="h-4 w-4" />
-              )}
-            </Button>
-          </div>
+          <DialogTitle className="flex items-center gap-2">
+            <Monitor className="h-5 w-5" />
+            System Information
+          </DialogTitle>
+          <DialogDescription>
+            View system and backend-specific environment and capabilities
+          </DialogDescription>
         </DialogHeader>
 
         <div className="flex-1 overflow-y-auto">
-          {loading && !systemInfo ? (
-            <div className="flex items-center justify-center py-12">
-              <Loader2 className="h-6 w-6 animate-spin text-gray-400" />
-              <span className="ml-2 text-gray-400">Loading system information...</span>
+          <div className="space-y-6">
+            {/* Llamactl Version Section - Always shown */}
+            <div className="space-y-3">
+              <h3 className="font-semibold">Llamactl Version</h3>
+              <div className="bg-gray-900 rounded-lg p-4">
+                <pre className="text-sm text-gray-300 whitespace-pre-wrap font-mono">
+                  {getAppVersion()}
+                </pre>
+              </div>
             </div>
-          ) : error ? (
-            <div className="flex items-center gap-2 p-4 bg-destructive/10 border border-destructive/20 rounded-lg">
-              <AlertCircle className="h-4 w-4 text-destructive" />
-              <span className="text-sm text-destructive">{error}</span>
-            </div>
-          ) : systemInfo ? (
-            <div className="space-y-6">
-              {/* Llamactl Version Section */}
-              <div className="space-y-3">
-                <h3 className="font-semibold">Llamactl Version</h3>
-                
-                <div className="bg-gray-900 rounded-lg p-4">
-                  <pre className="text-sm text-gray-300 whitespace-pre-wrap font-mono">
-                    {getAppVersion()}
-                  </pre>
-                </div>
-              </div>
 
-              {/* Llama Server Version Section */}
-              <div className="space-y-3">
-                <h3 className="font-semibold">Llama Server Version</h3>
-                
-                <div className="bg-gray-900 rounded-lg p-4">
-                  <div className="mb-2">
-                    <span className="text-sm text-gray-400">$ llama-server --version</span>
-                  </div>
-                  <pre className="text-sm text-gray-300 whitespace-pre-wrap font-mono">
-                    {systemInfo.version}
-                  </pre>
+            {/* Backend Selection Section */}
+            <div className="space-y-3">
+              <h3 className="font-semibold">Backend Information</h3>
+              <div className="flex items-center gap-3">
+                <div className="flex-1">
+                  <SelectInput
+                    id="backend-select"
+                    label=""
+                    value={selectedBackend}
+                    onChange={(value) => handleBackendChange(value || BackendType.LLAMA_CPP)}
+                    options={BACKEND_OPTIONS}
+                    className="text-sm"
+                  />
                 </div>
-              </div>
-
-              {/* Devices Section */}
-              <div className="space-y-3">
-                <div className="flex items-center gap-2">
-                  <h3 className="font-semibold">Available Devices</h3>
-                </div>
-                
-                <div className="bg-gray-900 rounded-lg p-4">
-                  <div className="mb-2">
-                    <span className="text-sm text-gray-400">$ llama-server --list-devices</span>
-                  </div>
-                  <pre className="text-sm text-gray-300 whitespace-pre-wrap font-mono">
-                    {systemInfo.devices}
-                  </pre>
-                </div>
-              </div>
-
-              {/* Help Section */}
-              <div className="space-y-3">
-                <Button
-                  variant="ghost"
-                  onClick={() => setShowHelp(!showHelp)}
-                  className="flex items-center gap-2 p-0 h-auto font-semibold"
-                >
-                  {showHelp ? (
-                    <ChevronDown className="h-4 w-4" />
-                  ) : (
-                    <ChevronRight className="h-4 w-4" />
-                  )}
-                  <HelpCircle className="h-4 w-4" />
-                  Command Line Options
-                </Button>
-                
-                {showHelp && (
-                  <div className="bg-gray-900 rounded-lg p-4">
-                    <div className="mb-2">
-                      <span className="text-sm text-gray-400">$ llama-server --help</span>
-                    </div>
-                    <pre className="text-sm text-gray-300 whitespace-pre-wrap font-mono max-h-64 overflow-y-auto">
-                      {systemInfo.help}
-                    </pre>
-                  </div>
+                {selectedBackend === BackendType.LLAMA_CPP && (
+                  <Button
+                    variant="outline"
+                    size="sm"
+                    onClick={() => void fetchBackendInfo(selectedBackend)}
+                    disabled={loading}
+                  >
+                    {loading ? (
+                      <Loader2 className="h-4 w-4 animate-spin" />
+                    ) : (
+                      <RefreshCw className="h-4 w-4" />
+                    )}
+                  </Button>
                 )}
               </div>
             </div>
-          ) : null}
+
+            {/* Backend-specific content */}
+            {renderBackendSpecificContent()}
+          </div>
         </div>
 
         <DialogFooter>

webui/src/components/__tests__/InstanceCard.test.tsx

@@ -3,6 +3,7 @@ import { render, screen } from '@testing-library/react'
 import userEvent from '@testing-library/user-event'
 import InstanceCard from '@/components/InstanceCard'
 import type { Instance } from '@/types/instance'
+import { BackendType } from '@/types/instance'
 
 // Mock the health hook since we're not testing health logic here
 vi.mock('@/hooks/useInstanceHealth', () => ({
@@ -17,14 +18,14 @@ describe('InstanceCard - Instance Actions and State', () => {
 
   const stoppedInstance: Instance = {
     name: 'test-instance',
-    running: false,
-    options: { model: 'test-model.gguf' }
+    status: 'stopped',
+    options: { backend_type: BackendType.LLAMA_CPP, backend_options: { model: 'test-model.gguf' } }
   }
 
   const runningInstance: Instance = {
     name: 'running-instance',
-    running: true,
-    options: { model: 'running-model.gguf' }
+    status: 'running',
+    options: { backend_type: BackendType.LLAMA_CPP, backend_options: { model: 'running-model.gguf' } }
   }
 
 beforeEach(() => {
@@ -101,7 +102,7 @@ afterEach(() => {
 
     it('opens logs dialog when logs button clicked', async () => {
       const user = userEvent.setup()
-      
+
       render(
         <InstanceCard
           instance={stoppedInstance}
@@ -112,9 +113,13 @@ afterEach(() => {
         />
       )
 
+      // First click "More actions" to reveal the logs button
+      const moreActionsButton = screen.getByTitle('More actions')
+      await user.click(moreActionsButton)
+
       const logsButton = screen.getByTitle('View logs')
       await user.click(logsButton)
-      
+
       // Should open logs dialog (we can verify this by checking if dialog title appears)
       expect(screen.getByText(`Logs: ${stoppedInstance.name}`)).toBeInTheDocument()
     })
@@ -124,7 +129,7 @@ afterEach(() => {
     it('shows confirmation dialog and calls deleteInstance when confirmed', async () => {
       const user = userEvent.setup()
       const confirmSpy = vi.spyOn(window, 'confirm').mockReturnValue(true)
-      
+
       render(
         <InstanceCard
           instance={stoppedInstance}
@@ -135,19 +140,23 @@ afterEach(() => {
         />
       )
 
+      // First click "More actions" to reveal the delete button
+      const moreActionsButton = screen.getByTitle('More actions')
+      await user.click(moreActionsButton)
+
       const deleteButton = screen.getByTitle('Delete instance')
       await user.click(deleteButton)
-      
+
       expect(confirmSpy).toHaveBeenCalledWith('Are you sure you want to delete instance "test-instance"?')
       expect(mockDeleteInstance).toHaveBeenCalledWith('test-instance')
-      
+
       confirmSpy.mockRestore()
     })
 
     it('does not call deleteInstance when confirmation cancelled', async () => {
       const user = userEvent.setup()
       const confirmSpy = vi.spyOn(window, 'confirm').mockReturnValue(false)
-      
+
       render(
         <InstanceCard
           instance={stoppedInstance}
@@ -158,18 +167,24 @@ afterEach(() => {
         />
       )
 
+      // First click "More actions" to reveal the delete button
+      const moreActionsButton = screen.getByTitle('More actions')
+      await user.click(moreActionsButton)
+
       const deleteButton = screen.getByTitle('Delete instance')
       await user.click(deleteButton)
-      
+
       expect(confirmSpy).toHaveBeenCalled()
       expect(mockDeleteInstance).not.toHaveBeenCalled()
-      
+
       confirmSpy.mockRestore()
     })
   })
 
   describe('Button State Based on Instance Status', () => {
-    it('disables start button and enables stop button for running instance', () => {
+    it('disables start button and enables stop button for running instance', async () => {
+      const user = userEvent.setup()
+
       render(
         <InstanceCard
           instance={runningInstance}
@@ -180,12 +195,19 @@ afterEach(() => {
         />
       )
 
-      expect(screen.getByTitle('Start instance')).toBeDisabled()
+      expect(screen.queryByTitle('Start instance')).not.toBeInTheDocument()
       expect(screen.getByTitle('Stop instance')).not.toBeDisabled()
+
+      // Expand more actions to access delete button
+      const moreActionsButton = screen.getByTitle('More actions')
+      await user.click(moreActionsButton)
+
       expect(screen.getByTitle('Delete instance')).toBeDisabled() // Can't delete running instance
     })
 
-    it('enables start button and disables stop button for stopped instance', () => {
+    it('enables start button and disables stop button for stopped instance', async () => {
+      const user = userEvent.setup()
+
       render(
         <InstanceCard
           instance={stoppedInstance}
@@ -197,11 +219,18 @@ afterEach(() => {
       )
 
       expect(screen.getByTitle('Start instance')).not.toBeDisabled()
-      expect(screen.getByTitle('Stop instance')).toBeDisabled()
+      expect(screen.queryByTitle('Stop instance')).not.toBeInTheDocument()
+
+      // Expand more actions to access delete button
+      const moreActionsButton = screen.getByTitle('More actions')
+      await user.click(moreActionsButton)
+
       expect(screen.getByTitle('Delete instance')).not.toBeDisabled() // Can delete stopped instance
     })
 
-    it('edit and logs buttons are always enabled', () => {
+    it('edit and logs buttons are always enabled', async () => {
+      const user = userEvent.setup()
+
       render(
         <InstanceCard
           instance={runningInstance}
@@ -213,6 +242,11 @@ afterEach(() => {
       )
 
       expect(screen.getByTitle('Edit instance')).not.toBeDisabled()
+
+      // Expand more actions to access logs button
+      const moreActionsButton = screen.getByTitle('More actions')
+      await user.click(moreActionsButton)
+
       expect(screen.getByTitle('View logs')).not.toBeDisabled()
     })
   })
@@ -267,7 +301,7 @@ afterEach(() => {
   describe('Integration with LogsModal', () => {
     it('passes correct props to LogsModal', async () => {
       const user = userEvent.setup()
-      
+
       render(
         <InstanceCard
           instance={runningInstance}
@@ -278,20 +312,24 @@ afterEach(() => {
         />
       )
 
+      // First click "More actions" to reveal the logs button
+      const moreActionsButton = screen.getByTitle('More actions')
+      await user.click(moreActionsButton)
+
       // Open logs dialog
       await user.click(screen.getByTitle('View logs'))
-      
+
       // Verify dialog opened with correct instance data
       expect(screen.getByText('Logs: running-instance')).toBeInTheDocument()
-      
+
       // Close dialog to test close functionality
       const closeButtons = screen.getAllByText('Close')
-      const dialogCloseButton = closeButtons.find(button => 
+      const dialogCloseButton = closeButtons.find(button =>
         button.closest('[data-slot="dialog-content"]')
       )
       expect(dialogCloseButton).toBeTruthy()
       await user.click(dialogCloseButton!)
-      
+
       // Modal should close
       expect(screen.queryByText('Logs: running-instance')).not.toBeInTheDocument()
     })
@@ -301,7 +339,7 @@ afterEach(() => {
     it('handles instance with minimal data', () => {
       const minimalInstance: Instance = {
         name: 'minimal',
-        running: false,
+        status: 'stopped',
         options: {}
       }
 
@@ -323,7 +361,7 @@ afterEach(() => {
     it('handles instance with undefined options', () => {
       const instanceWithoutOptions: Instance = {
         name: 'no-options',
-        running: true,
+        status: 'running',
         options: undefined
       }
 

webui/src/components/__tests__/InstanceList.test.tsx

@@ -5,6 +5,7 @@ import InstanceList from '@/components/InstanceList'
 import { InstancesProvider } from '@/contexts/InstancesContext'
 import { instancesApi } from '@/lib/api'
 import type { Instance } from '@/types/instance'
+import { BackendType } from '@/types/instance'
 import { AuthProvider } from '@/contexts/AuthContext'
 
 // Mock the API
@@ -44,9 +45,9 @@ describe('InstanceList - State Management and UI Logic', () => {
   const mockEditInstance = vi.fn()
 
   const mockInstances: Instance[] = [
-    { name: 'instance-1', running: false, options: { model: 'model1.gguf' } },
-    { name: 'instance-2', running: true, options: { model: 'model2.gguf' } },
-    { name: 'instance-3', running: false, options: { model: 'model3.gguf' } }
+    { name: 'instance-1', status: 'stopped', options: { backend_type: BackendType.LLAMA_CPP, backend_options: { model: 'model1.gguf' } } },
+    { name: 'instance-2', status: 'running', options: { backend_type: BackendType.LLAMA_CPP, backend_options: { model: 'model2.gguf' } } },
+    { name: 'instance-3', status: 'stopped', options: { backend_type: BackendType.LLAMA_CPP, backend_options: { model: 'model3.gguf' } } }
   ]
 
   const DUMMY_API_KEY = 'test-api-key-123'

webui/src/components/__tests__/InstanceModal.test.tsx

@@ -3,6 +3,7 @@ import { render, screen, waitFor } from '@testing-library/react'
 import userEvent from '@testing-library/user-event'
 import InstanceDialog from '@/components/InstanceDialog'
 import type { Instance } from '@/types/instance'
+import { BackendType } from '@/types/instance'
 
 describe('InstanceModal - Form Logic and Validation', () => {
   const mockOnSave = vi.fn()
@@ -91,6 +92,7 @@ afterEach(() => {
 
       expect(mockOnSave).toHaveBeenCalledWith('my-instance', {
         auto_restart: true, // Default value
+        backend_type: BackendType.LLAMA_CPP
       })
     })
 
@@ -134,10 +136,10 @@ afterEach(() => {
   describe('Edit Mode', () => {
     const mockInstance: Instance = {
       name: 'existing-instance',
-      running: false,
+      status: 'stopped',
       options: {
-        model: 'test-model.gguf',
-        gpu_layers: 10,
+        backend_type: BackendType.LLAMA_CPP,
+        backend_options: { model: 'test-model.gguf', gpu_layers: 10 },
         auto_restart: false
       }
     }
@@ -177,15 +179,15 @@ afterEach(() => {
       await user.click(screen.getByTestId('dialog-save-button'))
 
       expect(mockOnSave).toHaveBeenCalledWith('existing-instance', {
-        model: 'test-model.gguf',
-        gpu_layers: 10,
+        backend_type: BackendType.LLAMA_CPP,
+        backend_options: { model: 'test-model.gguf', gpu_layers: 10 },
         auto_restart: false
       })
     })
 
     it('shows correct button text for running vs stopped instances', () => {
-      const runningInstance: Instance = { ...mockInstance, running: true }
-      
+      const runningInstance: Instance = { ...mockInstance, status: 'running' }
+
       const { rerender } = render(
         <InstanceDialog
           open={true}
@@ -271,35 +273,13 @@ afterEach(() => {
 
       expect(mockOnSave).toHaveBeenCalledWith('test-instance', {
         auto_restart: true,
+        backend_type: BackendType.LLAMA_CPP,
         max_restarts: 5,
         restart_delay: 10
       })
     })
   })
 
-  describe('Advanced Fields Toggle', () => {
-    it('shows advanced fields when toggle clicked', async () => {
-      const user = userEvent.setup()
-      
-      render(
-        <InstanceDialog
-          open={true}
-          onOpenChange={mockOnOpenChange}
-          onSave={mockOnSave}
-        />
-      )
-
-      // Advanced fields should be hidden initially
-      expect(screen.queryByText(/Advanced Configuration/)).toBeInTheDocument()
-      
-      // Click to expand
-      await user.click(screen.getByText(/Advanced Configuration/))
-
-      // Should show more configuration options
-      // Note: Specific fields depend on zodFormUtils configuration
-      // We're testing the toggle behavior, not specific fields
-    })
-  })
 
   describe('Form Data Handling', () => {
     it('cleans up undefined values before submission', async () => {
@@ -321,6 +301,7 @@ afterEach(() => {
       // Should only include non-empty values
       expect(mockOnSave).toHaveBeenCalledWith('clean-instance', {
         auto_restart: true, // Only this default value should be included
+        backend_type: BackendType.LLAMA_CPP
       })
     })
 
@@ -345,7 +326,8 @@ afterEach(() => {
 
       expect(mockOnSave).toHaveBeenCalledWith('numeric-test', {
         auto_restart: true,
-        gpu_layers: 15, // Should be number, not string
+        backend_type: BackendType.LLAMA_CPP,
+        backend_options: { gpu_layers: 15 }, // Should be number, not string
       })
     })
   })

webui/src/components/form/ArrayInput.tsx

@@ -0,0 +1,62 @@
+import React from 'react'
+import { Input } from '@/components/ui/input'
+import { Label } from '@/components/ui/label'
+
+interface ArrayInputProps {
+  id: string
+  label: string
+  value: string[] | undefined
+  onChange: (value: string[] | undefined) => void
+  placeholder?: string
+  description?: string
+  disabled?: boolean
+  className?: string
+}
+
+const ArrayInput: React.FC<ArrayInputProps> = ({
+  id,
+  label,
+  value,
+  onChange,
+  placeholder = "item1, item2, item3",
+  description,
+  disabled = false,
+  className
+}) => {
+  const handleChange = (inputValue: string) => {
+    if (inputValue === '') {
+      onChange(undefined)
+      return
+    }
+
+    const arrayValue = inputValue
+      .split(',')
+      .map(s => s.trim())
+      .filter(Boolean)
+
+    onChange(arrayValue.length > 0 ? arrayValue : undefined)
+  }
+
+  return (
+    <div className="grid gap-2">
+      <Label htmlFor={id}>
+        {label}
+      </Label>
+      <Input
+        id={id}
+        type="text"
+        value={Array.isArray(value) ? value.join(', ') : ''}
+        onChange={(e) => handleChange(e.target.value)}
+        placeholder={placeholder}
+        disabled={disabled}
+        className={className}
+      />
+      {description && (
+        <p className="text-sm text-muted-foreground">{description}</p>
+      )}
+      <p className="text-xs text-muted-foreground">Separate multiple values with commas</p>
+    </div>
+  )
+}
+
+export default ArrayInput

webui/src/components/form/CheckboxInput.tsx

@@ -0,0 +1,42 @@
+import React from 'react'
+import { Checkbox } from '@/components/ui/checkbox'
+import { Label } from '@/components/ui/label'
+
+interface CheckboxInputProps {
+  id: string
+  label: string
+  value: boolean | undefined
+  onChange: (value: boolean) => void
+  description?: string
+  disabled?: boolean
+  className?: string
+}
+
+const CheckboxInput: React.FC<CheckboxInputProps> = ({
+  id,
+  label,
+  value,
+  onChange,
+  description,
+  disabled = false,
+  className
+}) => {
+  return (
+    <div className={`flex items-center space-x-2 ${className || ''}`}>
+      <Checkbox
+        id={id}
+        checked={value === true}
+        onCheckedChange={(checked) => onChange(!!checked)}
+        disabled={disabled}
+      />
+      <Label htmlFor={id} className="text-sm font-normal">
+        {label}
+        {description && (
+          <span className="text-muted-foreground ml-1">- {description}</span>
+        )}
+      </Label>
+    </div>
+  )
+}
+
+export default CheckboxInput

webui/src/components/form/EnvironmentVariablesInput.tsx

@@ -0,0 +1,144 @@
+import React, { useState } from 'react'
+import { Input } from '@/components/ui/input'
+import { Label } from '@/components/ui/label'
+import { Button } from '@/components/ui/button'
+import { X, Plus } from 'lucide-react'
+
+interface EnvironmentVariablesInputProps {
+  id: string
+  label: string
+  value: Record<string, string> | undefined
+  onChange: (value: Record<string, string> | undefined) => void
+  description?: string
+  disabled?: boolean
+  className?: string
+}
+
+interface EnvVar {
+  key: string
+  value: string
+}
+
+const EnvironmentVariablesInput: React.FC<EnvironmentVariablesInputProps> = ({
+  id,
+  label,
+  value,
+  onChange,
+  description,
+  disabled = false,
+  className
+}) => {
+  // Convert the value object to an array of key-value pairs for editing
+  const envVarsFromValue = value
+    ? Object.entries(value).map(([key, val]) => ({ key, value: val }))
+    : []
+
+  const [envVars, setEnvVars] = useState<EnvVar[]>(
+    envVarsFromValue.length > 0 ? envVarsFromValue : [{ key: '', value: '' }]
+  )
+
+  // Update parent component when env vars change
+  const updateParent = (newEnvVars: EnvVar[]) => {
+    // Filter out empty entries
+    const validVars = newEnvVars.filter(env => env.key.trim() !== '' && env.value.trim() !== '')
+
+    if (validVars.length === 0) {
+      onChange(undefined)
+    } else {
+      const envObject = validVars.reduce((acc, env) => {
+        acc[env.key.trim()] = env.value.trim()
+        return acc
+      }, {} as Record<string, string>)
+      onChange(envObject)
+    }
+  }
+
+  const handleKeyChange = (index: number, newKey: string) => {
+    const newEnvVars = [...envVars]
+    newEnvVars[index].key = newKey
+    setEnvVars(newEnvVars)
+    updateParent(newEnvVars)
+  }
+
+  const handleValueChange = (index: number, newValue: string) => {
+    const newEnvVars = [...envVars]
+    newEnvVars[index].value = newValue
+    setEnvVars(newEnvVars)
+    updateParent(newEnvVars)
+  }
+
+  const addEnvVar = () => {
+    const newEnvVars = [...envVars, { key: '', value: '' }]
+    setEnvVars(newEnvVars)
+  }
+
+  const removeEnvVar = (index: number) => {
+    if (envVars.length === 1) {
+      // Reset to empty if it's the last one
+      const newEnvVars = [{ key: '', value: '' }]
+      setEnvVars(newEnvVars)
+      updateParent(newEnvVars)
+    } else {
+      const newEnvVars = envVars.filter((_, i) => i !== index)
+      setEnvVars(newEnvVars)
+      updateParent(newEnvVars)
+    }
+  }
+
+  return (
+    <div className={`grid gap-2 ${className || ''}`}>
+      <Label htmlFor={id}>
+        {label}
+      </Label>
+      <div className="space-y-2">
+        {envVars.map((envVar, index) => (
+          <div key={index} className="flex gap-2 items-center">
+            <Input
+              placeholder="Variable name"
+              value={envVar.key}
+              onChange={(e) => handleKeyChange(index, e.target.value)}
+              disabled={disabled}
+              className="flex-1"
+            />
+            <Input
+              placeholder="Variable value"
+              value={envVar.value}
+              onChange={(e) => handleValueChange(index, e.target.value)}
+              disabled={disabled}
+              className="flex-1"
+            />
+            <Button
+              type="button"
+              variant="outline"
+              size="sm"
+              onClick={() => removeEnvVar(index)}
+              disabled={disabled}
+              className="shrink-0"
+            >
+              <X className="h-4 w-4" />
+            </Button>
+          </div>
+        ))}
+        <Button
+          type="button"
+          variant="outline"
+          size="sm"
+          onClick={addEnvVar}
+          disabled={disabled}
+          className="w-fit"
+        >
+          <Plus className="h-4 w-4 mr-2" />
+          Add Variable
+        </Button>
+      </div>
+      {description && (
+        <p className="text-sm text-muted-foreground">{description}</p>
+      )}
+      <p className="text-xs text-muted-foreground">
+        Environment variables that will be passed to the backend process
+      </p>
+    </div>
+  )
+}
+
+export default EnvironmentVariablesInput

webui/src/components/form/NumberInput.tsx

@@ -0,0 +1,60 @@
+import React from 'react'
+import { Input } from '@/components/ui/input'
+import { Label } from '@/components/ui/label'
+
+interface NumberInputProps {
+  id: string
+  label: string
+  value: number | undefined
+  onChange: (value: number | undefined) => void
+  placeholder?: string
+  description?: string
+  disabled?: boolean
+  className?: string
+}
+
+const NumberInput: React.FC<NumberInputProps> = ({
+  id,
+  label,
+  value,
+  onChange,
+  placeholder,
+  description,
+  disabled = false,
+  className
+}) => {
+  const handleChange = (inputValue: string) => {
+    if (inputValue === '') {
+      onChange(undefined)
+      return
+    }
+
+    const numValue = parseFloat(inputValue)
+    if (!isNaN(numValue)) {
+      onChange(numValue)
+    }
+  }
+
+  return (
+    <div className="grid gap-2">
+      <Label htmlFor={id}>
+        {label}
+      </Label>
+      <Input
+        id={id}
+        type="number"
+        step="any"
+        value={value !== undefined ? value : ''}
+        onChange={(e) => handleChange(e.target.value)}
+        placeholder={placeholder}
+        disabled={disabled}
+        className={className}
+      />
+      {description && (
+        <p className="text-sm text-muted-foreground">{description}</p>
+      )}
+    </div>
+  )
+}
+
+export default NumberInput

webui/src/components/form/SelectInput.tsx

@@ -0,0 +1,55 @@
+import React from 'react'
+import { Label } from '@/components/ui/label'
+
+interface SelectOption {
+  value: string
+  label: string
+}
+
+interface SelectInputProps {
+  id: string
+  label: string
+  value: string | undefined
+  onChange: (value: string | undefined) => void
+  options: SelectOption[]
+  description?: string
+  disabled?: boolean
+  className?: string
+}
+
+const SelectInput: React.FC<SelectInputProps> = ({
+  id,
+  label,
+  value,
+  onChange,
+  options,
+  description,
+  disabled = false,
+  className
+}) => {
+  return (
+    <div className="grid gap-2">
+      <Label htmlFor={id}>
+        {label}
+      </Label>
+      <select
+        id={id}
+        value={value || ''}
+        onChange={(e) => onChange(e.target.value || undefined)}
+        disabled={disabled}
+        className={`flex h-10 w-full rounded-md border border-input bg-background px-3 py-2 text-sm ring-offset-background focus-visible:outline-none focus-visible:ring-2 focus-visible:ring-ring focus-visible:ring-offset-2 disabled:cursor-not-allowed disabled:opacity-50 ${className || ''}`}
+      >
+        {options.map(option => (
+          <option key={option.value} value={option.value}>
+            {option.label}
+          </option>
+        ))}
+      </select>
+      {description && (
+        <p className="text-sm text-muted-foreground">{description}</p>
+      )}
+    </div>
+  )
+}
+
+export default SelectInput

webui/src/components/form/TextInput.tsx

@@ -0,0 +1,47 @@
+import React from 'react'
+import { Input } from '@/components/ui/input'
+import { Label } from '@/components/ui/label'
+
+interface TextInputProps {
+  id: string
+  label: string
+  value: string | number | undefined
+  onChange: (value: string | undefined) => void
+  placeholder?: string
+  description?: string
+  disabled?: boolean
+  className?: string
+}
+
+const TextInput: React.FC<TextInputProps> = ({
+  id,
+  label,
+  value,
+  onChange,
+  placeholder,
+  description,
+  disabled = false,
+  className
+}) => {
+  return (
+    <div className="grid gap-2">
+      <Label htmlFor={id}>
+        {label}
+      </Label>
+      <Input
+        id={id}
+        type="text"
+        value={typeof value === 'string' || typeof value === 'number' ? value : ''}
+        onChange={(e) => onChange(e.target.value || undefined)}
+        placeholder={placeholder}
+        disabled={disabled}
+        className={className}
+      />
+      {description && (
+        <p className="text-sm text-muted-foreground">{description}</p>
+      )}
+    </div>
+  )
+}
+
+export default TextInput

webui/src/components/instance/AutoRestartConfiguration.tsx

@@ -0,0 +1,53 @@
+import React from 'react'
+import type { CreateInstanceOptions } from '@/types/instance'
+import CheckboxInput from '@/components/form/CheckboxInput'
+import NumberInput from '@/components/form/NumberInput'
+
+interface AutoRestartConfigurationProps {
+  formData: CreateInstanceOptions
+  onChange: (key: keyof CreateInstanceOptions, value: any) => void
+}
+
+const AutoRestartConfiguration: React.FC<AutoRestartConfigurationProps> = ({
+  formData,
+  onChange
+}) => {
+  const isAutoRestartEnabled = formData.auto_restart === true
+
+  return (
+    <div className="space-y-4">
+      <h3 className="text-lg font-medium">Auto Restart Configuration</h3>
+
+      <CheckboxInput
+        id="auto_restart"
+        label="Auto Restart"
+        value={formData.auto_restart}
+        onChange={(value) => onChange('auto_restart', value)}
+        description="Automatically restart the instance on failure"
+      />
+
+      {isAutoRestartEnabled && (
+        <div className="ml-6 space-y-4 border-l-2 border-muted pl-4">
+          <NumberInput
+            id="max_restarts"
+            label="Max Restarts"
+            value={formData.max_restarts}
+            onChange={(value) => onChange('max_restarts', value)}
+            placeholder="3"
+            description="Maximum number of restart attempts (0 = unlimited)"
+          />
+          <NumberInput
+            id="restart_delay"
+            label="Restart Delay (seconds)"
+            value={formData.restart_delay}
+            onChange={(value) => onChange('restart_delay', value)}
+            placeholder="5"
+            description="Delay in seconds before attempting restart"
+          />
+        </div>
+      )}
+    </div>
+  )
+}
+
+export default AutoRestartConfiguration

webui/src/components/instance/BackendConfiguration.tsx

@@ -0,0 +1,54 @@
+import React from 'react'
+import type { CreateInstanceOptions } from '@/types/instance'
+import { getBasicBackendFields, getAdvancedBackendFields } from '@/lib/zodFormUtils'
+import BackendFormField from '@/components/BackendFormField'
+
+interface BackendConfigurationProps {
+  formData: CreateInstanceOptions
+  onBackendFieldChange: (key: string, value: any) => void
+  showAdvanced?: boolean
+}
+
+const BackendConfiguration: React.FC<BackendConfigurationProps> = ({
+  formData,
+  onBackendFieldChange,
+  showAdvanced = false
+}) => {
+  const basicBackendFields = getBasicBackendFields(formData.backend_type)
+  const advancedBackendFields = getAdvancedBackendFields(formData.backend_type)
+
+  return (
+    <div className="space-y-4">
+      <h3 className="text-lg font-medium">Backend Configuration</h3>
+
+      {/* Basic backend fields */}
+      {basicBackendFields.map((fieldKey) => (
+        <BackendFormField
+          key={fieldKey}
+          fieldKey={fieldKey}
+          value={(formData.backend_options as any)?.[fieldKey]}
+          onChange={onBackendFieldChange}
+        />
+      ))}
+
+      {/* Advanced backend fields */}
+      {showAdvanced && advancedBackendFields.length > 0 && (
+        <div className="space-y-4 pl-6 border-l-2 border-muted">
+          <h4 className="text-md font-medium">Advanced Backend Configuration</h4>
+          {advancedBackendFields
+            .sort()
+            .map((fieldKey) => (
+              <BackendFormField
+                key={fieldKey}
+                fieldKey={fieldKey}
+                value={(formData.backend_options as any)?.[fieldKey]}
+                onChange={onBackendFieldChange}
+              />
+            ))}
+        </div>
+      )}
+    </div>
+  )
+}
+
+export default BackendConfiguration

webui/src/components/instance/BackendConfigurationCard.tsx

@@ -0,0 +1,117 @@
+import React, { useState } from 'react'
+import { BackendType, type CreateInstanceOptions } from '@/types/instance'
+import { Card, CardContent, CardHeader, CardTitle } from '@/components/ui/card'
+import { Button } from '@/components/ui/button'
+import { Terminal, ChevronDown, ChevronRight } from 'lucide-react'
+import { getBasicBackendFields, getAdvancedBackendFields } from '@/lib/zodFormUtils'
+import BackendFormField from '@/components/BackendFormField'
+import SelectInput from '@/components/form/SelectInput'
+
+interface BackendConfigurationCardProps {
+  formData: CreateInstanceOptions
+  onBackendFieldChange: (key: string, value: unknown) => void
+  onChange: (key: keyof CreateInstanceOptions, value: unknown) => void
+  onParseCommand: () => void
+}
+
+const BackendConfigurationCard: React.FC<BackendConfigurationCardProps> = ({
+  formData,
+  onBackendFieldChange,
+  onChange,
+  onParseCommand
+}) => {
+  const [showAdvanced, setShowAdvanced] = useState(false)
+  const basicBackendFields = getBasicBackendFields(formData.backend_type)
+  const advancedBackendFields = getAdvancedBackendFields(formData.backend_type)
+
+  return (
+    <Card>
+      <CardHeader>
+        <CardTitle>Backend Configuration</CardTitle>
+      </CardHeader>
+      <CardContent className="space-y-6">
+        {/* Backend Type Selection */}
+        <SelectInput
+          id="backend_type"
+          label="Backend Type"
+          value={formData.backend_type || BackendType.LLAMA_CPP}
+          onChange={(value) => onChange('backend_type', value)}
+          options={[
+            { value: BackendType.LLAMA_CPP, label: 'Llama Server' },
+            { value: BackendType.MLX_LM, label: 'MLX LM' },
+            { value: BackendType.VLLM, label: 'vLLM' }
+          ]}
+          description="Select the backend server type"
+        />
+
+        {/* Parse Command Section */}
+        <div className="flex flex-col gap-2">
+          <Button
+            variant="outline"
+            onClick={onParseCommand}
+            className="flex items-center gap-2 w-fit"
+          >
+            <Terminal className="h-4 w-4" />
+            Parse Command
+          </Button>
+          <p className="text-sm text-muted-foreground">
+            Import settings from your backend command
+          </p>
+        </div>
+
+        {/* Basic Backend Options */}
+        {basicBackendFields.length > 0 && (
+          <div className="space-y-4">
+            <h3 className="text-md font-medium">Basic Backend Options</h3>
+            {basicBackendFields.map((fieldKey) => (
+              <BackendFormField
+                key={fieldKey}
+                fieldKey={fieldKey}
+                value={(formData.backend_options as Record<string, unknown>)?.[fieldKey] as string | number | boolean | string[] | undefined}
+                onChange={onBackendFieldChange}
+              />
+            ))}
+          </div>
+        )}
+
+        {/* Advanced Backend Options */}
+        {advancedBackendFields.length > 0 && (
+          <div className="space-y-4">
+            <Button
+              variant="ghost"
+              onClick={() => setShowAdvanced(!showAdvanced)}
+              className="flex items-center gap-2 p-0 h-auto font-medium"
+            >
+              {showAdvanced ? (
+                <ChevronDown className="h-4 w-4" />
+              ) : (
+                <ChevronRight className="h-4 w-4" />
+              )}
+              Advanced Backend Options
+              <span className="text-muted-foreground text-sm font-normal">
+                ({advancedBackendFields.length} options)
+              </span>
+            </Button>
+
+            {showAdvanced && (
+              <div className="space-y-4 pl-6 border-l-2 border-muted">
+                {advancedBackendFields
+                  .sort()
+                  .map((fieldKey) => (
+                    <BackendFormField
+                      key={fieldKey}
+                      fieldKey={fieldKey}
+                      value={(formData.backend_options as Record<string, unknown>)?.[fieldKey] as string | number | boolean | string[] | undefined}
+                      onChange={onBackendFieldChange}
+                    />
+                  ))}
+              </div>
+            )}
+          </div>
+        )}
+      </CardContent>
+    </Card>
+  )
+}
+
+export default BackendConfigurationCard

webui/src/components/instance/InstanceSettingsCard.tsx

@@ -0,0 +1,93 @@
+import React from 'react'
+import type { CreateInstanceOptions } from '@/types/instance'
+import { Card, CardContent, CardHeader, CardTitle } from '@/components/ui/card'
+import { Label } from '@/components/ui/label'
+import { Input } from '@/components/ui/input'
+import AutoRestartConfiguration from '@/components/instance/AutoRestartConfiguration'
+import NumberInput from '@/components/form/NumberInput'
+import CheckboxInput from '@/components/form/CheckboxInput'
+import EnvironmentVariablesInput from '@/components/form/EnvironmentVariablesInput'
+
+interface InstanceSettingsCardProps {
+  instanceName: string
+  nameError: string
+  isEditing: boolean
+  formData: CreateInstanceOptions
+  onNameChange: (name: string) => void
+  onChange: (key: keyof CreateInstanceOptions, value: unknown) => void
+}
+
+const InstanceSettingsCard: React.FC<InstanceSettingsCardProps> = ({
+  instanceName,
+  nameError,
+  isEditing,
+  formData,
+  onNameChange,
+  onChange
+}) => {
+  return (
+    <Card>
+      <CardHeader>
+        <CardTitle>Instance Settings</CardTitle>
+      </CardHeader>
+      <CardContent className="space-y-6">
+        {/* Instance Name */}
+        <div className="grid gap-2">
+          <Label htmlFor="name">
+            Instance Name <span className="text-red-500">*</span>
+          </Label>
+          <Input
+            id="name"
+            value={instanceName}
+            onChange={(e) => onNameChange(e.target.value)}
+            placeholder="my-instance"
+            disabled={isEditing}
+            className={nameError ? "border-red-500" : ""}
+          />
+          {nameError && <p className="text-sm text-red-500">{nameError}</p>}
+          <p className="text-sm text-muted-foreground">
+            Unique identifier for the instance
+          </p>
+        </div>
+
+        {/* Auto Restart Configuration */}
+        <AutoRestartConfiguration
+          formData={formData}
+          onChange={onChange}
+        />
+
+        {/* Basic Instance Options */}
+        <div className="space-y-4">
+          <h3 className="text-lg font-medium">Basic Instance Options</h3>
+
+          <NumberInput
+            id="idle_timeout"
+            label="Idle Timeout (minutes)"
+            value={formData.idle_timeout}
+            onChange={(value) => onChange('idle_timeout', value)}
+            placeholder="30"
+            description="Minutes before stopping an idle instance"
+          />
+
+          <CheckboxInput
+            id="on_demand_start"
+            label="On Demand Start"
+            value={formData.on_demand_start}
+            onChange={(value) => onChange('on_demand_start', value)}
+            description="Start instance only when needed"
+          />
+
+          <EnvironmentVariablesInput
+            id="environment"
+            label="Environment Variables"
+            value={formData.environment}
+            onChange={(value) => onChange('environment', value)}
+            description="Custom environment variables for the instance"
+          />
+        </div>
+      </CardContent>
+    </Card>
+  )
+}
+
+export default InstanceSettingsCard

webui/src/contexts/AuthContext.tsx

@@ -1,4 +1,4 @@
-import { type ReactNode, createContext, useContext, useState, useEffect, useCallback } from 'react'
+import { type ReactNode, createContext, useCallback, useContext, useEffect, useState } from 'react'
 
 interface AuthContextState {
   isAuthenticated: boolean
@@ -62,7 +62,7 @@ export const AuthProvider = ({ children }: AuthProviderProps) => {
   // Validate API key by making a test request
   const validateApiKey = async (key: string): Promise<boolean> => {
     try {
-      const response = await fetch('/api/v1/instances', {
+      const response = await fetch(document.baseURI + 'api/v1/instances', {
         headers: {
           'Authorization': `Bearer ${key}`,
           'Content-Type': 'application/json'

webui/src/contexts/InstancesContext.tsx

@@ -112,9 +112,9 @@ export const InstancesProvider = ({ children }: InstancesProviderProps) => {
     try {
       setError(null)
       await instancesApi.start(name)
-      
-      // Update only this instance's running status
-      updateInstanceInMap(name, { running: true })
+
+      // Update only this instance's status
+      updateInstanceInMap(name, { status: "running" })
     } catch (err) {
       setError(err instanceof Error ? err.message : 'Failed to start instance')
     }
@@ -124,9 +124,9 @@ export const InstancesProvider = ({ children }: InstancesProviderProps) => {
     try {
       setError(null)
       await instancesApi.stop(name)
-      
-      // Update only this instance's running status
-      updateInstanceInMap(name, { running: false })
+
+      // Update only this instance's status
+      updateInstanceInMap(name, { status: "stopped" })
     } catch (err) {
       setError(err instanceof Error ? err.message : 'Failed to stop instance')
     }
@@ -136,9 +136,9 @@ export const InstancesProvider = ({ children }: InstancesProviderProps) => {
     try {
       setError(null)
       await instancesApi.restart(name)
-      
-      // Update only this instance's running status
-      updateInstanceInMap(name, { running: true })
+
+      // Update only this instance's status
+      updateInstanceInMap(name, { status: "running" })
     } catch (err) {
       setError(err instanceof Error ? err.message : 'Failed to restart instance')
     }

webui/src/contexts/__tests__/InstancesContext.test.tsx

@@ -4,6 +4,7 @@ import type { ReactNode } from "react";
 import { InstancesProvider, useInstances } from "@/contexts/InstancesContext";
 import { instancesApi } from "@/lib/api";
 import type { Instance } from "@/types/instance";
+import { BackendType } from "@/types/instance";
 import { AuthProvider } from "../AuthContext";
 
 // Mock the API module
@@ -41,19 +42,19 @@ function TestComponent() {
       <div data-testid="instances-count">{instances.length}</div>
       {instances.map((instance) => (
         <div key={instance.name} data-testid={`instance-${instance.name}`}>
-          {instance.name}:{instance.running.toString()}
+          {instance.name}:{instance.status}
         </div>
       ))}
 
       {/* Action buttons for testing with specific instances */}
       <button
-        onClick={() => createInstance("new-instance", { model: "test.gguf" })}
+        onClick={() => createInstance("new-instance", { backend_type: BackendType.LLAMA_CPP, backend_options: { model: "test.gguf" } })}
         data-testid="create-instance"
       >
         Create Instance
       </button>
       <button
-        onClick={() => updateInstance("instance1", { model: "updated.gguf" })}
+        onClick={() => updateInstance("instance1", { backend_type: BackendType.LLAMA_CPP, backend_options: { model: "updated.gguf" } })}
         data-testid="update-instance"
       >
         Update Instance
@@ -99,8 +100,8 @@ function renderWithProvider(children: ReactNode) {
 
 describe("InstancesContext", () => {
   const mockInstances: Instance[] = [
-    { name: "instance1", running: true, options: { model: "model1.gguf" } },
-    { name: "instance2", running: false, options: { model: "model2.gguf" } },
+    { name: "instance1", status: "running", options: { backend_type: BackendType.LLAMA_CPP, backend_options: { model: "model1.gguf" } } },
+    { name: "instance2", status: "stopped", options: { backend_type: BackendType.LLAMA_CPP, backend_options: { model: "model2.gguf" } } },
   ];
 
   beforeEach(() => {
@@ -132,10 +133,10 @@ describe("InstancesContext", () => {
         expect(screen.getByTestId("loading")).toHaveTextContent("false");
         expect(screen.getByTestId("instances-count")).toHaveTextContent("2");
         expect(screen.getByTestId("instance-instance1")).toHaveTextContent(
-          "instance1:true"
+          "instance1:running"
         );
         expect(screen.getByTestId("instance-instance2")).toHaveTextContent(
-          "instance2:false"
+          "instance2:stopped"
         );
       });
     });
@@ -158,8 +159,8 @@ describe("InstancesContext", () => {
     it("creates instance and adds it to state", async () => {
       const newInstance: Instance = {
         name: "new-instance",
-        running: false,
-        options: { model: "test.gguf" },
+        status: "stopped",
+        options: { backend_type: BackendType.LLAMA_CPP, backend_options: { model: "test.gguf" } },
       };
       vi.mocked(instancesApi.create).mockResolvedValue(newInstance);
 
@@ -174,14 +175,15 @@ describe("InstancesContext", () => {
 
       await waitFor(() => {
         expect(instancesApi.create).toHaveBeenCalledWith("new-instance", {
-          model: "test.gguf",
+          backend_type: BackendType.LLAMA_CPP,
+          backend_options: { model: "test.gguf" }
         });
       });
 
       await waitFor(() => {
         expect(screen.getByTestId("instances-count")).toHaveTextContent("3");
         expect(screen.getByTestId("instance-new-instance")).toHaveTextContent(
-          "new-instance:false"
+          "new-instance:stopped"
         );
       });
     });
@@ -214,8 +216,8 @@ describe("InstancesContext", () => {
     it("updates instance and maintains it in state", async () => {
       const updatedInstance: Instance = {
         name: "instance1",
-        running: true,
-        options: { model: "updated.gguf" },
+        status: "running",
+        options: { backend_type: BackendType.LLAMA_CPP, backend_options: { model: "updated.gguf" } },
       };
       vi.mocked(instancesApi.update).mockResolvedValue(updatedInstance);
 
@@ -230,7 +232,8 @@ describe("InstancesContext", () => {
 
       await waitFor(() => {
         expect(instancesApi.update).toHaveBeenCalledWith("instance1", {
-          model: "updated.gguf",
+          backend_type: BackendType.LLAMA_CPP,
+          backend_options: { model: "updated.gguf" }
         });
       });
 
@@ -251,7 +254,7 @@ describe("InstancesContext", () => {
         expect(screen.getByTestId("loading")).toHaveTextContent("false");
         // instance2 starts as not running
         expect(screen.getByTestId("instance-instance2")).toHaveTextContent(
-          "instance2:false"
+          "instance2:stopped"
         );
       });
 
@@ -262,7 +265,7 @@ describe("InstancesContext", () => {
         expect(instancesApi.start).toHaveBeenCalledWith("instance2");
         // The running state should be updated to true
         expect(screen.getByTestId("instance-instance2")).toHaveTextContent(
-          "instance2:true"
+          "instance2:running"
         );
       });
     });
@@ -276,7 +279,7 @@ describe("InstancesContext", () => {
         expect(screen.getByTestId("loading")).toHaveTextContent("false");
         // instance1 starts as running
         expect(screen.getByTestId("instance-instance1")).toHaveTextContent(
-          "instance1:true"
+          "instance1:running"
         );
       });
 
@@ -287,7 +290,7 @@ describe("InstancesContext", () => {
         expect(instancesApi.stop).toHaveBeenCalledWith("instance1");
         // The running state should be updated to false
         expect(screen.getByTestId("instance-instance1")).toHaveTextContent(
-          "instance1:false"
+          "instance1:stopped"
         );
       });
     });
@@ -383,7 +386,7 @@ describe("InstancesContext", () => {
       // Test that operations don't interfere with each other
       const newInstance: Instance = {
         name: "new-instance",
-        running: false,
+        status: "stopped",
         options: {},
       };
       vi.mocked(instancesApi.create).mockResolvedValue(newInstance);
@@ -411,7 +414,7 @@ describe("InstancesContext", () => {
         expect(screen.getByTestId("instances-count")).toHaveTextContent("3"); // Still 3
         // But the running state should change
         expect(screen.getByTestId("instance-instance2")).toHaveTextContent(
-          "instance2:true"
+          "instance2:running"
         );
       });
     });

webui/src/hooks/useInstanceHealth.ts

@@ -1,14 +1,19 @@
 // ui/src/hooks/useInstanceHealth.ts
 import { useState, useEffect } from 'react'
-import type { HealthStatus } from '@/types/instance'
+import type { HealthStatus, InstanceStatus } from '@/types/instance'
 import { healthService } from '@/lib/healthService'
 
-export function useInstanceHealth(instanceName: string, isRunning: boolean): HealthStatus | undefined {
+export function useInstanceHealth(instanceName: string, instanceStatus: InstanceStatus): HealthStatus | undefined {
   const [health, setHealth] = useState<HealthStatus | undefined>()
 
   useEffect(() => {
-    if (!isRunning) {
-      setHealth(undefined)
+    if (instanceStatus === "stopped") {
+      setHealth({ status: "unknown", lastChecked: new Date() })
+      return
+    }
+
+    if (instanceStatus === "failed") {
+      setHealth({ status: instanceStatus, lastChecked: new Date() })
       return
     }
 
@@ -17,9 +22,9 @@ export function useInstanceHealth(instanceName: string, isRunning: boolean): Hea
       setHealth(healthStatus)
     })
 
-    // Cleanup subscription on unmount or when running changes
+    // Cleanup subscription on unmount or when instanceStatus changes
     return unsubscribe
-  }, [instanceName, isRunning])
+  }, [instanceName, instanceStatus])
 
   return health
 }

webui/src/lib/__tests__/api.test.ts

@@ -1,5 +1,5 @@
-import { describe, it, expect, vi, beforeEach } from 'vitest'
 import { instancesApi } from '@/lib/api'
+import { beforeEach, describe, expect, it, vi } from 'vitest'
 
 // Mock fetch globally
 const mockFetch = vi.fn()
@@ -11,11 +11,13 @@ describe('API Error Handling', () => {
   })
 
   it('converts HTTP errors to meaningful messages', async () => {
-    mockFetch.mockResolvedValue({
+    const mockResponse = {
       ok: false,
       status: 409,
-      text: () => Promise.resolve('Instance already exists')
-    })
+      text: () => Promise.resolve('Instance already exists'),
+      clone: function() { return this }
+    }
+    mockFetch.mockResolvedValue(mockResponse)
 
     await expect(instancesApi.create('existing', {}))
       .rejects
@@ -23,11 +25,13 @@ describe('API Error Handling', () => {
   })
 
   it('handles empty error responses gracefully', async () => {
-    mockFetch.mockResolvedValue({
+    const mockResponse = {
       ok: false,
       status: 500,
-      text: () => Promise.resolve('')
-    })
+      text: () => Promise.resolve(''),
+      clone: function() { return this }
+    }
+    mockFetch.mockResolvedValue(mockResponse)
 
     await expect(instancesApi.list())
       .rejects
@@ -53,7 +57,9 @@ describe('API Error Handling', () => {
     await instancesApi.getLogs('test-instance', 100)
 
     expect(mockFetch).toHaveBeenCalledWith(
-      '/api/v1/instances/test-instance/logs?lines=100',
+      expect.stringMatching(
+        /^https?:\/\/[^/]+\/api\/v1\/instances\/test-instance\/logs\?lines=100$/
+      ),
       expect.any(Object)
     )
   })

webui/src/lib/api.ts

@@ -1,6 +1,10 @@
 import type { CreateInstanceOptions, Instance } from "@/types/instance";
+import { handleApiError } from "./errorUtils";
 
-const API_BASE = "/api/v1";
+// Adding baseURI as a prefix to support being served behind a subpath
+// e.g. when llmamctl's `/` is served behind a reverse proxy at `/proxy/...`
+// the baseURI will be `/proxy/` and the API calls will be made to `/proxy/api/v1/<endpoint>`
+export const API_BASE = document.baseURI + "api/v1";
 
 // Base API call function with error handling
 async function apiCall<T>(
@@ -30,25 +34,8 @@ async function apiCall<T>(
       headers,
     });
 
-    // Handle authentication errors
-    if (response.status === 401) {
-      throw new Error('Authentication required');
-    }
-
-    if (!response.ok) {
-      // Try to get error message from response
-      let errorMessage = `HTTP ${response.status}`;
-      try {
-        const errorText = await response.text();
-        if (errorText) {
-          errorMessage += `: ${errorText}`;
-        }
-      } catch {
-        // If we can't read the error, just use status
-      }
-
-      throw new Error(errorMessage);
-    }
+    // Handle errors using centralized error handler
+    await handleApiError(response);
 
     // Handle empty responses (like DELETE)
     if (response.status === 204) {
@@ -60,6 +47,11 @@ async function apiCall<T>(
       const text = await response.text();
       return text as T;
     } else {
+      // Handle empty responses for JSON endpoints
+      const contentLength = response.headers.get('content-length');
+      if (contentLength === '0') {
+        return {} as T; // Return empty object for empty JSON responses
+      }
       const data = await response.json() as T;
       return data;
     }
@@ -71,16 +63,44 @@ async function apiCall<T>(
   }
 }
 
-// Server API functions
+// Server API functions (moved to llama-cpp backend)
 export const serverApi = {
-  // GET /server/help
-  getHelp: () => apiCall<string>("/server/help", {}, "text"),
+  // GET /backends/llama-cpp/help
+  getHelp: () => apiCall<string>("/backends/llama-cpp/help", {}, "text"),
 
-  // GET /server/version
-  getVersion: () => apiCall<string>("/server/version", {}, "text"),
+  // GET /backends/llama-cpp/version
+  getVersion: () => apiCall<string>("/backends/llama-cpp/version", {}, "text"),
 
-  // GET /server/devices
-  getDevices: () => apiCall<string>("/server/devices", {}, "text"),
+  // GET /backends/llama-cpp/devices
+  getDevices: () => apiCall<string>("/backends/llama-cpp/devices", {}, "text"),
+};
+
+// Backend API functions
+export const backendsApi = {
+  llamaCpp: {
+    // POST /backends/llama-cpp/parse-command
+    parseCommand: (command: string) =>
+      apiCall<CreateInstanceOptions>('/backends/llama-cpp/parse-command', {
+        method: 'POST',
+        body: JSON.stringify({ command }),
+      }),
+  },
+  mlx: {
+    // POST /backends/mlx/parse-command
+    parseCommand: (command: string) =>
+      apiCall<CreateInstanceOptions>('/backends/mlx/parse-command', {
+        method: 'POST',
+        body: JSON.stringify({ command }),
+      }),
+  },
+  vllm: {
+    // POST /backends/vllm/parse-command
+    parseCommand: (command: string) =>
+      apiCall<CreateInstanceOptions>('/backends/vllm/parse-command', {
+        method: 'POST',
+        body: JSON.stringify({ command }),
+      }),
+  },
 };
 
 // Instance API functions
@@ -136,5 +156,5 @@ export const instancesApi = {
   },
 
   // GET /instances/{name}/proxy/health
-  getHealth: (name: string) => apiCall<any>(`/instances/${name}/proxy/health`),
+  getHealth: (name: string) => apiCall<Record<string, unknown>>(`/instances/${name}/proxy/health`),
 };

webui/src/lib/errorUtils.ts

@@ -0,0 +1,33 @@
+/**
+ * Parses error response from API calls and returns a formatted error message
+ */
+export async function parseErrorResponse(response: Response): Promise<string> {
+  let errorMessage = `HTTP ${response.status}`
+
+  try {
+    const errorText = await response.text()
+    if (errorText) {
+      errorMessage += `: ${errorText}`
+    }
+  } catch {
+    // If we can't read the error, just use status
+  }
+
+  return errorMessage
+}
+
+/**
+ * Handles common API call errors and throws appropriate Error objects
+ */
+export async function handleApiError(response: Response): Promise<void> {
+  // Handle authentication errors
+  if (response.status === 401) {
+    throw new Error('Authentication required')
+  }
+
+  if (!response.ok) {
+    // Clone the response before reading to avoid consuming the body stream
+    const errorMessage = await parseErrorResponse(response.clone())
+    throw new Error(errorMessage)
+  }
+}

webui/src/lib/zodFormUtils.ts

@@ -1,35 +1,24 @@
-import { type CreateInstanceOptions, getAllFieldKeys } from '@/schemas/instanceOptions'
+import {
+  type LlamaCppBackendOptions,
+  type MlxBackendOptions,
+  type VllmBackendOptions,
+  LlamaCppBackendOptionsSchema,
+  MlxBackendOptionsSchema,
+  VllmBackendOptionsSchema,
+  getAllLlamaCppFieldKeys,
+  getAllMlxFieldKeys,
+  getAllVllmFieldKeys,
+  getLlamaCppFieldType,
+  getMlxFieldType,
+  getVllmFieldType
+} from '@/schemas/instanceOptions'
 
-// Only define the basic fields we want to show by default
-export const basicFieldsConfig: Record<string, {
+// LlamaCpp backend-specific basic fields
+const basicLlamaCppFieldsConfig: Record<string, {
   label: string
   description?: string
   placeholder?: string
-  required?: boolean
 }> = {
-  auto_restart: {
-    label: 'Auto Restart',
-    description: 'Automatically restart the instance on failure'
-  },
-  max_restarts: {
-    label: 'Max Restarts',
-    placeholder: '3',
-    description: 'Maximum number of restart attempts (0 = unlimited)'
-  },
-  restart_delay: {
-    label: 'Restart Delay (seconds)',
-    placeholder: '5',
-    description: 'Delay in seconds before attempting restart'
-  },
-  idle_timeout: {
-    label: 'Idle Timeout (minutes)',
-    placeholder: '60',
-    description: 'Time in minutes before instance is considered idle and stopped'
-  },
-  on_demand_start: {
-    label: 'On-Demand Start',
-    description: 'Start instance upon receiving OpenAI-compatible API request'
-  },
   model: {
     label: 'Model Path',
     placeholder: '/path/to/model.gguf',
@@ -52,17 +41,135 @@ export const basicFieldsConfig: Record<string, {
   }
 }
 
-export function isBasicField(key: keyof CreateInstanceOptions): boolean {
-  return key in basicFieldsConfig
+// MLX backend-specific basic fields
+const basicMlxFieldsConfig: Record<string, {
+  label: string
+  description?: string
+  placeholder?: string
+}> = {
+  model: {
+    label: 'Model',
+    placeholder: 'mlx-community/Mistral-7B-Instruct-v0.3-4bit',
+    description: 'The path to the MLX model weights, tokenizer, and config'
+  },
+  temp: {
+    label: 'Temperature',
+    placeholder: '0.0',
+    description: 'Default sampling temperature (default: 0.0)'
+  },
+  top_p: {
+    label: 'Top-P',
+    placeholder: '1.0',
+    description: 'Default nucleus sampling top-p (default: 1.0)'
+  },
+  top_k: {
+    label: 'Top-K',
+    placeholder: '0',
+    description: 'Default top-k sampling (default: 0, disables top-k)'
+  },
+  min_p: {
+    label: 'Min-P',
+    placeholder: '0.0',
+    description: 'Default min-p sampling (default: 0.0, disables min-p)'
+  },
+  max_tokens: {
+    label: 'Max Tokens',
+    placeholder: '512',
+    description: 'Default maximum number of tokens to generate (default: 512)'
+  }
 }
 
-export function getBasicFields(): (keyof CreateInstanceOptions)[] {
-  return Object.keys(basicFieldsConfig) as (keyof CreateInstanceOptions)[]
+// vLLM backend-specific basic fields
+const basicVllmFieldsConfig: Record<string, {
+  label: string
+  description?: string
+  placeholder?: string
+}> = {
+  model: {
+    label: 'Model',
+    placeholder: 'microsoft/DialoGPT-medium',
+    description: 'The name or path of the Hugging Face model to use'
+  },
+  tensor_parallel_size: {
+    label: 'Tensor Parallel Size',
+    placeholder: '1',
+    description: 'Number of GPUs to use for distributed serving'
+  },
+  gpu_memory_utilization: {
+    label: 'GPU Memory Utilization',
+    placeholder: '0.9',
+    description: 'The fraction of GPU memory to be used for the model executor'
+  }
 }
 
-export function getAdvancedFields(): (keyof CreateInstanceOptions)[] {
-  return getAllFieldKeys().filter(key => !isBasicField(key))
+// Backend field configuration lookup
+const backendFieldConfigs = {
+  mlx_lm: basicMlxFieldsConfig,
+  vllm: basicVllmFieldsConfig,
+  llama_cpp: basicLlamaCppFieldsConfig,
+} as const
+
+const backendFieldGetters = {
+  mlx_lm: getAllMlxFieldKeys,
+  vllm: getAllVllmFieldKeys,
+  llama_cpp: getAllLlamaCppFieldKeys,
+} as const
+
+export function getBasicBackendFields(backendType?: string): string[] {
+  const normalizedType = (backendType || 'llama_cpp') as keyof typeof backendFieldConfigs
+  const config = backendFieldConfigs[normalizedType] || basicLlamaCppFieldsConfig
+  return Object.keys(config)
+}
+
+export function getAdvancedBackendFields(backendType?: string): string[] {
+  const normalizedType = (backendType || 'llama_cpp') as keyof typeof backendFieldGetters
+  const fieldGetter = backendFieldGetters[normalizedType] || getAllLlamaCppFieldKeys
+  const basicConfig = backendFieldConfigs[normalizedType] || basicLlamaCppFieldsConfig
+
+  return fieldGetter().filter(key => !(key in basicConfig))
+}
+
+// Combined backend fields config for use in BackendFormField
+export const basicBackendFieldsConfig: Record<string, {
+  label: string
+  description?: string
+  placeholder?: string
+}> = {
+  ...basicLlamaCppFieldsConfig,
+  ...basicMlxFieldsConfig,
+  ...basicVllmFieldsConfig
+}
+
+// Get field type for any backend option (union type)
+export function getBackendFieldType(key: string): 'text' | 'number' | 'boolean' | 'array' {
+  // Try to get type from LlamaCpp schema first
+  try {
+    if (LlamaCppBackendOptionsSchema.shape && key in LlamaCppBackendOptionsSchema.shape) {
+      return getLlamaCppFieldType(key as keyof LlamaCppBackendOptions)
+    }
+  } catch {
+    // Schema might not be available
+  }
+
+  // Try MLX schema
+  try {
+    if (MlxBackendOptionsSchema.shape && key in MlxBackendOptionsSchema.shape) {
+      return getMlxFieldType(key as keyof MlxBackendOptions)
+    }
+  } catch {
+    // Schema might not be available
+  }
+
+  // Try vLLM schema
+  try {
+    if (VllmBackendOptionsSchema.shape && key in VllmBackendOptionsSchema.shape) {
+      return getVllmFieldType(key as keyof VllmBackendOptions)
+    }
+  } catch {
+    // Schema might not be available
+  }
+
+  // Default fallback
+  return 'text'
 }
 
-// Re-export the Zod-based functions
-export { getFieldType } from '@/schemas/instanceOptions'

webui/src/schemas/backends/index.ts

@@ -0,0 +1,4 @@
+// Re-export all backend schemas from one place
+export * from './llamacpp'
+export * from './mlx'
+export * from './vllm'

webui/src/schemas/backends/llamacpp.ts

@@ -0,0 +1,192 @@
+import { z } from 'zod'
+
+// Define the LlamaCpp backend options schema
+export const LlamaCppBackendOptionsSchema = z.object({
+  // Common params
+  verbose_prompt: z.boolean().optional(),
+  threads: z.number().optional(),
+  threads_batch: z.number().optional(),
+  cpu_mask: z.string().optional(),
+  cpu_range: z.string().optional(),
+  cpu_strict: z.number().optional(),
+  prio: z.number().optional(),
+  poll: z.number().optional(),
+  cpu_mask_batch: z.string().optional(),
+  cpu_range_batch: z.string().optional(),
+  cpu_strict_batch: z.number().optional(),
+  prio_batch: z.number().optional(),
+  poll_batch: z.number().optional(),
+  ctx_size: z.number().optional(),
+  predict: z.number().optional(),
+  batch_size: z.number().optional(),
+  ubatch_size: z.number().optional(),
+  keep: z.number().optional(),
+  flash_attn: z.boolean().optional(),
+  no_perf: z.boolean().optional(),
+  escape: z.boolean().optional(),
+  no_escape: z.boolean().optional(),
+  rope_scaling: z.string().optional(),
+  rope_scale: z.number().optional(),
+  rope_freq_base: z.number().optional(),
+  rope_freq_scale: z.number().optional(),
+  yarn_orig_ctx: z.number().optional(),
+  yarn_ext_factor: z.number().optional(),
+  yarn_attn_factor: z.number().optional(),
+  yarn_beta_slow: z.number().optional(),
+  yarn_beta_fast: z.number().optional(),
+  dump_kv_cache: z.boolean().optional(),
+  no_kv_offload: z.boolean().optional(),
+  cache_type_k: z.string().optional(),
+  cache_type_v: z.string().optional(),
+  defrag_thold: z.number().optional(),
+  parallel: z.number().optional(),
+  mlock: z.boolean().optional(),
+  no_mmap: z.boolean().optional(),
+  numa: z.string().optional(),
+  device: z.string().optional(),
+  override_tensor: z.array(z.string()).optional(),
+  gpu_layers: z.number().optional(),
+  split_mode: z.string().optional(),
+  tensor_split: z.string().optional(),
+  main_gpu: z.number().optional(),
+  check_tensors: z.boolean().optional(),
+  override_kv: z.array(z.string()).optional(),
+  lora: z.array(z.string()).optional(),
+  lora_scaled: z.array(z.string()).optional(),
+  control_vector: z.array(z.string()).optional(),
+  control_vector_scaled: z.array(z.string()).optional(),
+  control_vector_layer_range: z.string().optional(),
+  model: z.string().optional(),
+  model_url: z.string().optional(),
+  hf_repo: z.string().optional(),
+  hf_repo_draft: z.string().optional(),
+  hf_file: z.string().optional(),
+  hf_repo_v: z.string().optional(),
+  hf_file_v: z.string().optional(),
+  hf_token: z.string().optional(),
+  log_disable: z.boolean().optional(),
+  log_file: z.string().optional(),
+  log_colors: z.boolean().optional(),
+  verbose: z.boolean().optional(),
+  verbosity: z.number().optional(),
+  log_prefix: z.boolean().optional(),
+  log_timestamps: z.boolean().optional(),
+
+  // Sampling params
+  samplers: z.string().optional(),
+  seed: z.number().optional(),
+  sampling_seq: z.string().optional(),
+  ignore_eos: z.boolean().optional(),
+  temp: z.number().optional(),
+  top_k: z.number().optional(),
+  top_p: z.number().optional(),
+  min_p: z.number().optional(),
+  xtc_probability: z.number().optional(),
+  xtc_threshold: z.number().optional(),
+  typical: z.number().optional(),
+  repeat_last_n: z.number().optional(),
+  repeat_penalty: z.number().optional(),
+  presence_penalty: z.number().optional(),
+  frequency_penalty: z.number().optional(),
+  dry_multiplier: z.number().optional(),
+  dry_base: z.number().optional(),
+  dry_allowed_length: z.number().optional(),
+  dry_penalty_last_n: z.number().optional(),
+  dry_sequence_breaker: z.array(z.string()).optional(),
+  dynatemp_range: z.number().optional(),
+  dynatemp_exp: z.number().optional(),
+  mirostat: z.number().optional(),
+  mirostat_lr: z.number().optional(),
+  mirostat_ent: z.number().optional(),
+  logit_bias: z.array(z.string()).optional(),
+  grammar: z.string().optional(),
+  grammar_file: z.string().optional(),
+  json_schema: z.string().optional(),
+  json_schema_file: z.string().optional(),
+
+  // Example-specific params
+  no_context_shift: z.boolean().optional(),
+  special: z.boolean().optional(),
+  no_warmup: z.boolean().optional(),
+  spm_infill: z.boolean().optional(),
+  pooling: z.string().optional(),
+  cont_batching: z.boolean().optional(),
+  no_cont_batching: z.boolean().optional(),
+  mmproj: z.string().optional(),
+  mmproj_url: z.string().optional(),
+  no_mmproj: z.boolean().optional(),
+  no_mmproj_offload: z.boolean().optional(),
+  alias: z.string().optional(),
+  host: z.string().optional(),
+  port: z.number().optional(),
+  path: z.string().optional(),
+  no_webui: z.boolean().optional(),
+  embedding: z.boolean().optional(),
+  reranking: z.boolean().optional(),
+  api_key: z.string().optional(),
+  api_key_file: z.string().optional(),
+  ssl_key_file: z.string().optional(),
+  ssl_cert_file: z.string().optional(),
+  chat_template_kwargs: z.string().optional(),
+  timeout: z.number().optional(),
+  threads_http: z.number().optional(),
+  cache_reuse: z.number().optional(),
+  metrics: z.boolean().optional(),
+  slots: z.boolean().optional(),
+  props: z.boolean().optional(),
+  no_slots: z.boolean().optional(),
+  slot_save_path: z.string().optional(),
+  jinja: z.boolean().optional(),
+  reasoning_format: z.string().optional(),
+  reasoning_budget: z.number().optional(),
+  chat_template: z.string().optional(),
+  chat_template_file: z.string().optional(),
+  no_prefill_assistant: z.boolean().optional(),
+  slot_prompt_similarity: z.number().optional(),
+  lora_init_without_apply: z.boolean().optional(),
+  draft_max: z.number().optional(),
+  draft_min: z.number().optional(),
+  draft_p_min: z.number().optional(),
+  ctx_size_draft: z.number().optional(),
+  device_draft: z.string().optional(),
+  gpu_layers_draft: z.number().optional(),
+  model_draft: z.string().optional(),
+  cache_type_k_draft: z.string().optional(),
+  cache_type_v_draft: z.string().optional(),
+
+  // Audio/TTS params
+  model_vocoder: z.string().optional(),
+  tts_use_guide_tokens: z.boolean().optional(),
+
+  // Default model params
+  embd_bge_small_en_default: z.boolean().optional(),
+  embd_e5_small_en_default: z.boolean().optional(),
+  embd_gte_small_default: z.boolean().optional(),
+  fim_qwen_1_5b_default: z.boolean().optional(),
+  fim_qwen_3b_default: z.boolean().optional(),
+  fim_qwen_7b_default: z.boolean().optional(),
+  fim_qwen_7b_spec: z.boolean().optional(),
+  fim_qwen_14b_spec: z.boolean().optional(),
+})
+
+// Infer the TypeScript type from the schema
+export type LlamaCppBackendOptions = z.infer<typeof LlamaCppBackendOptionsSchema>
+
+// Helper to get all LlamaCpp backend option field keys
+export function getAllLlamaCppFieldKeys(): (keyof LlamaCppBackendOptions)[] {
+  return Object.keys(LlamaCppBackendOptionsSchema.shape) as (keyof LlamaCppBackendOptions)[]
+}
+
+// Get field type for LlamaCpp backend options
+export function getLlamaCppFieldType(key: keyof LlamaCppBackendOptions): 'text' | 'number' | 'boolean' | 'array' {
+  const fieldSchema = LlamaCppBackendOptionsSchema.shape[key]
+  if (!fieldSchema) return 'text'
+
+  // Handle ZodOptional wrapper
+  const innerSchema = fieldSchema instanceof z.ZodOptional ? fieldSchema.unwrap() : fieldSchema
+
+  if (innerSchema instanceof z.ZodBoolean) return 'boolean'
+  if (innerSchema instanceof z.ZodNumber) return 'number'
+  if (innerSchema instanceof z.ZodArray) return 'array'
+  return 'text' // ZodString and others default to text
+}

webui/src/schemas/backends/mlx.ts

@@ -0,0 +1,51 @@
+import { z } from 'zod'
+
+// Define the MLX backend options schema
+export const MlxBackendOptionsSchema = z.object({
+  // Basic connection options
+  model: z.string().optional(),
+  host: z.string().optional(),
+  port: z.number().optional(),
+
+  // Model and adapter options
+  adapter_path: z.string().optional(),
+  draft_model: z.string().optional(),
+  num_draft_tokens: z.number().optional(),
+  trust_remote_code: z.boolean().optional(),
+
+  // Logging and templates
+  log_level: z.enum(['DEBUG', 'INFO', 'WARNING', 'ERROR', 'CRITICAL']).optional(),
+  chat_template: z.string().optional(),
+  use_default_chat_template: z.boolean().optional(),
+  chat_template_args: z.string().optional(), // JSON string
+
+  // Sampling defaults
+  temp: z.number().optional(),     // Note: MLX uses "temp" not "temperature"
+  top_p: z.number().optional(),
+  top_k: z.number().optional(),
+  min_p: z.number().optional(),
+  max_tokens: z.number().optional(),
+})
+
+// Infer the TypeScript type from the schema
+export type MlxBackendOptions = z.infer<typeof MlxBackendOptionsSchema>
+
+// Helper to get all MLX backend option field keys
+export function getAllMlxFieldKeys(): (keyof MlxBackendOptions)[] {
+  return Object.keys(MlxBackendOptionsSchema.shape) as (keyof MlxBackendOptions)[]
+}
+
+// Get field type for MLX backend options
+export function getMlxFieldType(key: keyof MlxBackendOptions): 'text' | 'number' | 'boolean' | 'array' {
+  const fieldSchema = MlxBackendOptionsSchema.shape[key]
+  if (!fieldSchema) return 'text'
+
+  // Handle ZodOptional wrapper
+  const innerSchema = fieldSchema instanceof z.ZodOptional ? fieldSchema.unwrap() : fieldSchema
+
+  if (innerSchema instanceof z.ZodBoolean) return 'boolean'
+  if (innerSchema instanceof z.ZodNumber) return 'number'
+  if (innerSchema instanceof z.ZodArray) return 'array'
+  if (innerSchema instanceof z.ZodEnum) return 'text' // Enum treated as text/select
+  return 'text' // ZodString and others default to text
+}

webui/src/schemas/backends/vllm.ts

@@ -0,0 +1,150 @@
+import { z } from 'zod'
+
+// Define the vLLM backend options schema
+export const VllmBackendOptionsSchema = z.object({
+  // Basic connection options (auto-assigned by llamactl)
+  host: z.string().optional(),
+  port: z.number().optional(),
+
+  // Model and engine configuration
+  model: z.string().optional(),
+  tokenizer: z.string().optional(),
+  skip_tokenizer_init: z.boolean().optional(),
+  revision: z.string().optional(),
+  code_revision: z.string().optional(),
+  tokenizer_revision: z.string().optional(),
+  tokenizer_mode: z.string().optional(),
+  trust_remote_code: z.boolean().optional(),
+  download_dir: z.string().optional(),
+  load_format: z.string().optional(),
+  config_format: z.string().optional(),
+  dtype: z.string().optional(),
+  kv_cache_dtype: z.string().optional(),
+  quantization_param_path: z.string().optional(),
+  seed: z.number().optional(),
+  max_model_len: z.number().optional(),
+  guided_decoding_backend: z.string().optional(),
+  distributed_executor_backend: z.string().optional(),
+  worker_use_ray: z.boolean().optional(),
+  ray_workers_use_nsight: z.boolean().optional(),
+
+  // Performance and serving configuration
+  block_size: z.number().optional(),
+  enable_prefix_caching: z.boolean().optional(),
+  disable_sliding_window: z.boolean().optional(),
+  use_v2_block_manager: z.boolean().optional(),
+  num_lookahead_slots: z.number().optional(),
+  swap_space: z.number().optional(),
+  cpu_offload_gb: z.number().optional(),
+  gpu_memory_utilization: z.number().optional(),
+  num_gpu_blocks_override: z.number().optional(),
+  max_num_batched_tokens: z.number().optional(),
+  max_num_seqs: z.number().optional(),
+  max_logprobs: z.number().optional(),
+  disable_log_stats: z.boolean().optional(),
+  quantization: z.string().optional(),
+  rope_scaling: z.string().optional(),
+  rope_theta: z.number().optional(),
+  enforce_eager: z.boolean().optional(),
+  max_context_len_to_capture: z.number().optional(),
+  max_seq_len_to_capture: z.number().optional(),
+  disable_custom_all_reduce: z.boolean().optional(),
+  tokenizer_pool_size: z.number().optional(),
+  tokenizer_pool_type: z.string().optional(),
+  tokenizer_pool_extra_config: z.string().optional(),
+  enable_lora_bias: z.boolean().optional(),
+  lora_extra_vocab_size: z.number().optional(),
+  lora_rank: z.number().optional(),
+  prompt_lookback_distance: z.number().optional(),
+  preemption_mode: z.string().optional(),
+
+  // Distributed and parallel processing
+  tensor_parallel_size: z.number().optional(),
+  pipeline_parallel_size: z.number().optional(),
+  max_parallel_loading_workers: z.number().optional(),
+  disable_async_output_proc: z.boolean().optional(),
+  worker_class: z.string().optional(),
+  enabled_lora_modules: z.string().optional(),
+  max_lora_rank: z.number().optional(),
+  fully_sharded_loras: z.boolean().optional(),
+  lora_modules: z.string().optional(),
+  prompt_adapters: z.string().optional(),
+  max_prompt_adapter_token: z.number().optional(),
+  device: z.string().optional(),
+  scheduler_delay: z.number().optional(),
+  enable_chunked_prefill: z.boolean().optional(),
+  speculative_model: z.string().optional(),
+  speculative_model_quantization: z.string().optional(),
+  speculative_revision: z.string().optional(),
+  speculative_max_model_len: z.number().optional(),
+  speculative_disable_by_batch_size: z.number().optional(),
+  ngpt_speculative_length: z.number().optional(),
+  speculative_disable_mqa: z.boolean().optional(),
+  model_loader_extra_config: z.string().optional(),
+  ignore_patterns: z.string().optional(),
+  preloaded_lora_modules: z.string().optional(),
+
+  // OpenAI server specific options
+  uds: z.string().optional(),
+  uvicorn_log_level: z.string().optional(),
+  response_role: z.string().optional(),
+  ssl_keyfile: z.string().optional(),
+  ssl_certfile: z.string().optional(),
+  ssl_ca_certs: z.string().optional(),
+  ssl_cert_reqs: z.number().optional(),
+  root_path: z.string().optional(),
+  middleware: z.array(z.string()).optional(),
+  return_tokens_as_token_ids: z.boolean().optional(),
+  disable_frontend_multiprocessing: z.boolean().optional(),
+  enable_auto_tool_choice: z.boolean().optional(),
+  tool_call_parser: z.string().optional(),
+  tool_server: z.string().optional(),
+  chat_template: z.string().optional(),
+  chat_template_content_format: z.string().optional(),
+  allow_credentials: z.boolean().optional(),
+  allowed_origins: z.array(z.string()).optional(),
+  allowed_methods: z.array(z.string()).optional(),
+  allowed_headers: z.array(z.string()).optional(),
+  api_key: z.array(z.string()).optional(),
+  enable_log_outputs: z.boolean().optional(),
+  enable_token_usage: z.boolean().optional(),
+  enable_async_engine_debug: z.boolean().optional(),
+  engine_use_ray: z.boolean().optional(),
+  disable_log_requests: z.boolean().optional(),
+  max_log_len: z.number().optional(),
+
+  // Additional engine configuration
+  task: z.string().optional(),
+  multi_modal_config: z.string().optional(),
+  limit_mm_per_prompt: z.string().optional(),
+  enable_sleep_mode: z.boolean().optional(),
+  enable_chunking_request: z.boolean().optional(),
+  compilation_config: z.string().optional(),
+  disable_sliding_window_mask: z.boolean().optional(),
+  enable_trtllm_engine_latency: z.boolean().optional(),
+  override_pooling_config: z.string().optional(),
+  override_neuron_config: z.string().optional(),
+  override_kv_cache_align_size: z.number().optional(),
+})
+
+// Infer the TypeScript type from the schema
+export type VllmBackendOptions = z.infer<typeof VllmBackendOptionsSchema>
+
+// Helper to get all vLLM backend option field keys
+export function getAllVllmFieldKeys(): (keyof VllmBackendOptions)[] {
+  return Object.keys(VllmBackendOptionsSchema.shape) as (keyof VllmBackendOptions)[]
+}
+
+// Get field type for vLLM backend options
+export function getVllmFieldType(key: keyof VllmBackendOptions): 'text' | 'number' | 'boolean' | 'array' {
+  const fieldSchema = VllmBackendOptionsSchema.shape[key]
+  if (!fieldSchema) return 'text'
+
+  // Handle ZodOptional wrapper
+  const innerSchema = fieldSchema instanceof z.ZodOptional ? fieldSchema.unwrap() : fieldSchema
+
+  if (innerSchema instanceof z.ZodBoolean) return 'boolean'
+  if (innerSchema instanceof z.ZodNumber) return 'number'
+  if (innerSchema instanceof z.ZodArray) return 'array'
+  return 'text' // ZodString and others default to text
+}

webui/src/schemas/instanceOptions.ts

@@ -1,6 +1,30 @@
+import { BackendType } from '@/types/instance'
 import { z } from 'zod'
 
-// Define the Zod schema
+// Import backend schemas from separate files
+import {
+  LlamaCppBackendOptionsSchema,
+  type LlamaCppBackendOptions,
+  getAllLlamaCppFieldKeys,
+  getLlamaCppFieldType,
+  MlxBackendOptionsSchema,
+  type MlxBackendOptions,
+  getAllMlxFieldKeys,
+  getMlxFieldType,
+  VllmBackendOptionsSchema,
+  type VllmBackendOptions,
+  getAllVllmFieldKeys,
+  getVllmFieldType
+} from './backends'
+
+// Backend options union
+export const BackendOptionsSchema = z.union([
+  LlamaCppBackendOptionsSchema,
+  MlxBackendOptionsSchema,
+  VllmBackendOptionsSchema,
+])
+
+// Define the main create instance options schema
 export const CreateInstanceOptionsSchema = z.object({
   // Restart options
   auto_restart: z.boolean().optional(),
@@ -9,191 +33,51 @@ export const CreateInstanceOptionsSchema = z.object({
   idle_timeout: z.number().optional(),
   on_demand_start: z.boolean().optional(),
 
-  // Common params
-  verbose_prompt: z.boolean().optional(),
-  threads: z.number().optional(),
-  threads_batch: z.number().optional(),
-  cpu_mask: z.string().optional(),
-  cpu_range: z.string().optional(),
-  cpu_strict: z.number().optional(),
-  prio: z.number().optional(),
-  poll: z.number().optional(),
-  cpu_mask_batch: z.string().optional(),
-  cpu_range_batch: z.string().optional(),
-  cpu_strict_batch: z.number().optional(),
-  prio_batch: z.number().optional(),
-  poll_batch: z.number().optional(),
-  ctx_size: z.number().optional(),
-  predict: z.number().optional(),
-  batch_size: z.number().optional(),
-  ubatch_size: z.number().optional(),
-  keep: z.number().optional(),
-  flash_attn: z.boolean().optional(),
-  no_perf: z.boolean().optional(),
-  escape: z.boolean().optional(),
-  no_escape: z.boolean().optional(),
-  rope_scaling: z.string().optional(),
-  rope_scale: z.number().optional(),
-  rope_freq_base: z.number().optional(),
-  rope_freq_scale: z.number().optional(),
-  yarn_orig_ctx: z.number().optional(),
-  yarn_ext_factor: z.number().optional(),
-  yarn_attn_factor: z.number().optional(),
-  yarn_beta_slow: z.number().optional(),
-  yarn_beta_fast: z.number().optional(),
-  dump_kv_cache: z.boolean().optional(),
-  no_kv_offload: z.boolean().optional(),
-  cache_type_k: z.string().optional(),
-  cache_type_v: z.string().optional(),
-  defrag_thold: z.number().optional(),
-  parallel: z.number().optional(),
-  mlock: z.boolean().optional(),
-  no_mmap: z.boolean().optional(),
-  numa: z.string().optional(),
-  device: z.string().optional(),
-  override_tensor: z.array(z.string()).optional(),
-  gpu_layers: z.number().optional(),
-  split_mode: z.string().optional(),
-  tensor_split: z.string().optional(),
-  main_gpu: z.number().optional(),
-  check_tensors: z.boolean().optional(),
-  override_kv: z.array(z.string()).optional(),
-  lora: z.array(z.string()).optional(),
-  lora_scaled: z.array(z.string()).optional(),
-  control_vector: z.array(z.string()).optional(),
-  control_vector_scaled: z.array(z.string()).optional(),
-  control_vector_layer_range: z.string().optional(),
-  model: z.string().optional(),
-  model_url: z.string().optional(),
-  hf_repo: z.string().optional(),
-  hf_repo_draft: z.string().optional(),
-  hf_file: z.string().optional(),
-  hf_repo_v: z.string().optional(),
-  hf_file_v: z.string().optional(),
-  hf_token: z.string().optional(),
-  log_disable: z.boolean().optional(),
-  log_file: z.string().optional(),
-  log_colors: z.boolean().optional(),
-  verbose: z.boolean().optional(),
-  verbosity: z.number().optional(),
-  log_prefix: z.boolean().optional(),
-  log_timestamps: z.boolean().optional(),
+  // Environment variables
+  environment: z.record(z.string(), z.string()).optional(),
 
-  // Sampling params
-  samplers: z.string().optional(),
-  seed: z.number().optional(),
-  sampling_seq: z.string().optional(),
-  ignore_eos: z.boolean().optional(),
-  temp: z.number().optional(),
-  top_k: z.number().optional(),
-  top_p: z.number().optional(),
-  min_p: z.number().optional(),
-  xtc_probability: z.number().optional(),
-  xtc_threshold: z.number().optional(),
-  typical: z.number().optional(),
-  repeat_last_n: z.number().optional(),
-  repeat_penalty: z.number().optional(),
-  presence_penalty: z.number().optional(),
-  frequency_penalty: z.number().optional(),
-  dry_multiplier: z.number().optional(),
-  dry_base: z.number().optional(),
-  dry_allowed_length: z.number().optional(),
-  dry_penalty_last_n: z.number().optional(),
-  dry_sequence_breaker: z.array(z.string()).optional(),
-  dynatemp_range: z.number().optional(),
-  dynatemp_exp: z.number().optional(),
-  mirostat: z.number().optional(),
-  mirostat_lr: z.number().optional(),
-  mirostat_ent: z.number().optional(),
-  logit_bias: z.array(z.string()).optional(),
-  grammar: z.string().optional(),
-  grammar_file: z.string().optional(),
-  json_schema: z.string().optional(),
-  json_schema_file: z.string().optional(),
-
-  // Example-specific params
-  no_context_shift: z.boolean().optional(),
-  special: z.boolean().optional(),
-  no_warmup: z.boolean().optional(),
-  spm_infill: z.boolean().optional(),
-  pooling: z.string().optional(),
-  cont_batching: z.boolean().optional(),
-  no_cont_batching: z.boolean().optional(),
-  mmproj: z.string().optional(),
-  mmproj_url: z.string().optional(),
-  no_mmproj: z.boolean().optional(),
-  no_mmproj_offload: z.boolean().optional(),
-  alias: z.string().optional(),
-  host: z.string().optional(),
-  port: z.number().optional(),
-  path: z.string().optional(),
-  no_webui: z.boolean().optional(),
-  embedding: z.boolean().optional(),
-  reranking: z.boolean().optional(),
-  api_key: z.string().optional(),
-  api_key_file: z.string().optional(),
-  ssl_key_file: z.string().optional(),
-  ssl_cert_file: z.string().optional(),
-  chat_template_kwargs: z.string().optional(),
-  timeout: z.number().optional(),
-  threads_http: z.number().optional(),
-  cache_reuse: z.number().optional(),
-  metrics: z.boolean().optional(),
-  slots: z.boolean().optional(),
-  props: z.boolean().optional(),
-  no_slots: z.boolean().optional(),
-  slot_save_path: z.string().optional(),
-  jinja: z.boolean().optional(),
-  reasoning_format: z.string().optional(),
-  reasoning_budget: z.number().optional(),
-  chat_template: z.string().optional(),
-  chat_template_file: z.string().optional(),
-  no_prefill_assistant: z.boolean().optional(),
-  slot_prompt_similarity: z.number().optional(),
-  lora_init_without_apply: z.boolean().optional(),
-  draft_max: z.number().optional(),
-  draft_min: z.number().optional(),
-  draft_p_min: z.number().optional(),
-  ctx_size_draft: z.number().optional(),
-  device_draft: z.string().optional(),
-  gpu_layers_draft: z.number().optional(),
-  model_draft: z.string().optional(),
-  cache_type_k_draft: z.string().optional(),
-  cache_type_v_draft: z.string().optional(),
-
-  // Audio/TTS params
-  model_vocoder: z.string().optional(),
-  tts_use_guide_tokens: z.boolean().optional(),
-
-  // Default model params
-  embd_bge_small_en_default: z.boolean().optional(),
-  embd_e5_small_en_default: z.boolean().optional(),
-  embd_gte_small_default: z.boolean().optional(),
-  fim_qwen_1_5b_default: z.boolean().optional(),
-  fim_qwen_3b_default: z.boolean().optional(),
-  fim_qwen_7b_default: z.boolean().optional(),
-  fim_qwen_7b_spec: z.boolean().optional(),
-  fim_qwen_14b_spec: z.boolean().optional(),
+  // Backend configuration
+  backend_type: z.enum([BackendType.LLAMA_CPP, BackendType.MLX_LM, BackendType.VLLM]).optional(),
+  backend_options: BackendOptionsSchema.optional(),
 })
 
-// Infer the TypeScript type from the schema
+// Re-export types and schemas from backend files
+export {
+  LlamaCppBackendOptionsSchema,
+  MlxBackendOptionsSchema,
+  VllmBackendOptionsSchema,
+  type LlamaCppBackendOptions,
+  type MlxBackendOptions,
+  type VllmBackendOptions,
+  getAllLlamaCppFieldKeys,
+  getAllMlxFieldKeys,
+  getAllVllmFieldKeys,
+  getLlamaCppFieldType,
+  getMlxFieldType,
+  getVllmFieldType
+}
+
+// Infer the TypeScript types from the schemas
+export type BackendOptions = z.infer<typeof BackendOptionsSchema>
 export type CreateInstanceOptions = z.infer<typeof CreateInstanceOptionsSchema>
 
-// Helper to get all field keys
+// Helper to get all field keys for CreateInstanceOptions
 export function getAllFieldKeys(): (keyof CreateInstanceOptions)[] {
   return Object.keys(CreateInstanceOptionsSchema.shape) as (keyof CreateInstanceOptions)[]
 }
 
 // Get field type from Zod schema
-export function getFieldType(key: keyof CreateInstanceOptions): 'text' | 'number' | 'boolean' | 'array' {
+export function getFieldType(key: keyof CreateInstanceOptions): 'text' | 'number' | 'boolean' | 'array' | 'object' {
   const fieldSchema = CreateInstanceOptionsSchema.shape[key]
   if (!fieldSchema) return 'text'
-  
+
   // Handle ZodOptional wrapper
   const innerSchema = fieldSchema instanceof z.ZodOptional ? fieldSchema.unwrap() : fieldSchema
-  
+
   if (innerSchema instanceof z.ZodBoolean) return 'boolean'
   if (innerSchema instanceof z.ZodNumber) return 'number'
   if (innerSchema instanceof z.ZodArray) return 'array'
+  if (innerSchema instanceof z.ZodObject) return 'object'
+  if (innerSchema instanceof z.ZodRecord) return 'object' // Handle ZodRecord as object
   return 'text' // ZodString and others default to text
 }

webui/src/types/instance.ts

@@ -2,14 +2,26 @@ import type { CreateInstanceOptions } from '@/schemas/instanceOptions'
 
 export { type CreateInstanceOptions } from '@/schemas/instanceOptions'
 
+export const BackendType = {
+  LLAMA_CPP: 'llama_cpp',
+  MLX_LM: 'mlx_lm',
+  VLLM: 'vllm',
+  // MLX_VLM: 'mlx_vlm',  // Future expansion
+} as const
+
+export type BackendTypeValue = typeof BackendType[keyof typeof BackendType]
+
+export type InstanceStatus = 'running' | 'stopped' | 'failed'
+
 export interface HealthStatus {
-  status: 'ok' | 'loading' | 'error' | 'unknown'
+  status: 'ok' | 'loading' | 'error' | 'unknown' | 'failed'
   message?: string
   lastChecked: Date
 }
 
 export interface Instance {
   name: string;
-  running: boolean;
+  status: InstanceStatus;
   options?: CreateInstanceOptions;
+  docker_enabled?: boolean; // indicates backend is running via Docker
 }

webui/vite.config.ts

@@ -21,4 +21,6 @@ export default defineConfig({
     setupFiles: ['./src/test/setup.ts'],
     css: true,
   },
+  // ensures relative asset paths to support being served behind a subpath
+  base: "./"
 })