Merge pull request #23 from lordmathis/feat/start-on-request

feat: On-Demand Instance Start

.github/workflows/release.yaml

@@ -29,6 +29,8 @@ jobs:
           npm ci
 
       - name: Build Web UI
+        env:
+          VITE_APP_VERSION: ${{ github.ref_name }}
         run: |
           cd webui
           npm run build

CNAME

@@ -0,0 +1 @@
+llamactl.org

CONTRIBUTING.md

@@ -0,0 +1,138 @@
+# Contributing to Llamactl
+
+Thank you for considering contributing to Llamactl! This document outlines the development setup and contribution process.
+
+## Development Setup
+
+### Prerequisites
+
+- Go 1.24 or later
+- Node.js 22 or later
+- `llama-server` executable (from [llama.cpp](https://github.com/ggml-org/llama.cpp))
+
+### Getting Started
+
+1. **Clone the repository**
+   ```bash
+   git clone https://github.com/lordmathis/llamactl.git
+   cd llamactl
+   ```
+
+2. **Install dependencies**
+   ```bash
+   # Go dependencies
+   go mod download
+   
+   # Frontend dependencies
+   cd webui && npm ci && cd ..
+   ```
+
+3. **Run for development**
+   ```bash
+   # Start backend server
+   go run ./cmd/server
+   ```
+   Server will be available at `http://localhost:8080`
+   
+   ```bash
+   # In a separate terminal, start frontend dev server
+   cd webui && npm run dev
+   ```
+   Development UI will be available at `http://localhost:5173`
+
+4. **Common development commands**
+   ```bash
+   # Backend
+   go test ./... -v                    # Run tests
+   go test -race ./... -v              # Run with race detector
+   go fmt ./... && go vet ./...        # Format and vet code
+   
+   # Frontend (run from webui/ directory)
+   npm run test:run                    # Run tests once
+   npm run test                        # Run tests in watch mode
+   npm run type-check                  # TypeScript type checking
+   npm run lint:fix                    # Lint and fix issues
+   ```
+
+## Before Submitting a Pull Request
+
+### Required Checks
+
+All the following must pass:
+
+1. **Backend**
+   ```bash
+   go test ./... -v
+   go test -race ./... -v
+   go fmt ./... && go vet ./...
+   go build -o llamactl ./cmd/server
+   ```
+
+2. **Frontend**
+   ```bash
+   cd webui
+   npm run test:run
+   npm run type-check
+   npm run build
+   ```
+
+### API Documentation
+
+If changes affect API endpoints, update Swagger documentation:
+
+```bash
+# Install swag if needed
+go install github.com/swaggo/swag/cmd/swag@latest
+
+# Update Swagger comments in pkg/server/handlers.go
+# Then regenerate docs
+swag init -g cmd/server/main.go -o apidocs
+```
+
+## Pull Request Guidelines
+
+### Pull Request Titles
+Use this format for pull request titles:
+- `feat:` for new features
+- `fix:` for bug fixes  
+- `docs:` for documentation changes
+- `test:` for test additions or modifications
+- `refactor:` for code refactoring
+
+### Submission Process
+1. Create a feature branch from `main`
+2. Make changes following the coding standards
+3. Run all required checks listed above
+4. Update documentation if necessary
+5. Submit pull request with:
+   - Clear description of changes
+   - Reference to any related issues
+   - Screenshots for UI changes
+
+## Code Style and Testing
+
+### Testing Strategy
+- Backend tests use Go's built-in testing framework
+- Frontend tests use Vitest and React Testing Library
+- Run tests frequently during development
+- Add tests for new features and bug fixes
+
+### Go
+- Follow standard Go formatting (`go fmt`)
+- Use meaningful variable and function names  
+- Add comments for exported functions and types
+- Handle errors appropriately
+
+### TypeScript/React
+- Use TypeScript strictly (avoid `any` when possible)
+- Follow React hooks best practices
+- Use meaningful component and variable names
+- Prefer functional components over class components
+
+## Getting Help
+
+- Check existing [issues](https://github.com/lordmathis/llamactl/issues)
+- Review the [README.md](README.md) for usage documentation  
+- Look at existing code for patterns and conventions
+
+Thank you for contributing to Llamactl!

README.md

@@ -2,90 +2,140 @@
 
 ![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)
 
-A control server for managing multiple Llama Server instances with a web-based dashboard.
+**Management server and proxy for multiple llama.cpp instances with OpenAI-compatible API routing.**
 
-## Features
+## Why llamactl?
 
-- **Multi-instance Management**: Create, start, stop, restart, and delete multiple llama-server instances
+🚀 **Multiple Model Serving**: Run different models simultaneously (7B for speed, 70B for quality)  
-- **Web Dashboard**: Modern React-based UI for managing instances
+🔗 **OpenAI API Compatible**: Drop-in replacement - route requests by model name  
-- **Auto-restart**: Configurable automatic restart on instance failure
+🌐 **Web Dashboard**: Modern React UI for visual management (unlike CLI-only tools)  
-- **Instance Monitoring**: Real-time health checks and status monitoring
+🔐 **API Key Authentication**: Separate keys for management vs inference access  
-- **Log Management**: View, search, and download instance logs
+📊 **Instance Monitoring**: Health checks, auto-restart, log management  
-- **Data Persistence**: Persistent storage of instance state.
+⏳ **Idle Timeout Management**: Automatically stop idle instances after a configurable period  
-- **REST API**: Full API for programmatic control
+💡 **On-Demand Instance Start**: Automatically launch instances upon receiving OpenAI-compatible API requests  
-- **OpenAI Compatible**: Route requests to instances by instance name
+💾 **State Persistence**: Ensure instances remain intact across server restarts  
-- **Configuration Management**: Comprehensive llama-server parameter support
-- **System Information**: View llama-server version, devices, and help
-- **API Key Authentication**: Secure access with separate management and inference keys
 
-## Prerequisites
+![Dashboard Screenshot](docs/images/screenshot.png)
 
-This project requires `llama-server` from llama.cpp to be installed and available in your PATH.
+**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
 
-**Install llama.cpp:**
+## Quick Start
-Follow the installation instructions at https://github.com/ggml-org/llama.cpp
+
+```bash
+# 1. Install llama-server (one-time setup)
+# See: https://github.com/ggml-org/llama.cpp#quick-start
+
+# 2. Download and run llamactl
+LATEST_VERSION=$(curl -s https://api.github.com/repos/lordmathis/llamactl/releases/latest | grep '"tag_name":' | sed -E 's/.*"([^"]+)".*/\1/')
+curl -L https://github.com/lordmathis/llamactl/releases/download/${LATEST_VERSION}/llamactl-${LATEST_VERSION}-linux-amd64.tar.gz | tar -xz
+sudo mv llamactl /usr/local/bin/
+
+# 3. Start the server
+llamactl
+# Access dashboard at http://localhost:8080
+```
+
+## Usage
+
+### 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
+
+### Or use the REST API:
+```bash
+# Create 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}'
+
+# Use with OpenAI SDK
+curl -X POST localhost:8080/v1/chat/completions \
+  -H "Authorization: Bearer your-key" \
+  -d '{"model": "my-7b-model", "messages": [{"role": "user", "content": "Hello!"}]}'
+```
 
 ## Installation
 
-### Download Prebuilt Binaries
+### Option 1: Download Binary (Recommended)
 
-The easiest way to install llamactl is to download a prebuilt binary from the [releases page](https://github.com/lordmathis/llamactl/releases).
-
-**Linux/macOS:**
 ```bash
-# Download the latest release for your platform
+# Linux/macOS - Get latest version and download
-curl -L https://github.com/lordmathis/llamactl/releases/latest/download/llamactl-$(curl -s https://api.github.com/repos/lordmathis/llamactl/releases/latest | grep tag_name | cut -d '"' -f 4)-linux-amd64.tar.gz | tar -xz
+LATEST_VERSION=$(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
-# Move to PATH
 sudo mv llamactl /usr/local/bin/
 
-# Run the server
+# Or download manually from the releases page:
-llamactl
+# https://github.com/lordmathis/llamactl/releases/latest
+
+# Windows - Download from releases page
 ```
 
-**Manual Download:**
+### Option 2: Build from Source
-1. Go to the [releases page](https://github.com/lordmathis/llamactl/releases)
+Requires Go 1.24+ and Node.js 22+
-2. Download the appropriate archive for your platform
-3. Extract the archive and move the binary to a directory in your PATH
-
-### Build from Source
-
-If you prefer to build from source or need the latest development version:
-
-#### Build Requirements
-
-- Go 1.24 or later
-- Node.js 22 or later (for building the web UI)
-
-#### Building with Web UI
-
 ```bash
-# Clone the repository
 git clone https://github.com/lordmathis/llamactl.git
 cd llamactl
-
+cd webui && npm ci && npm run build && cd ..
-# Install Node.js dependencies
-cd webui
-npm ci
-
-# Build the web UI
-npm run build
-
-# Return to project root and build
-cd ..
 go build -o llamactl ./cmd/server
+```
 
-# Run the server
+## Prerequisites
-./llamactl
+
+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
 ```
 
 ## Configuration
 
-llamactl can be configured via configuration files or environment variables. Configuration is loaded in the following order of precedence:
+llamactl works out of the box with sensible defaults.
 
-1. Hardcoded defaults
+```yaml
-2. Configuration file
+server:
-3. Environment variables
+  host: "0.0.0.0"                # Server host to bind to
+  port: 8080                     # Server port to bind to
+  allowed_origins: ["*"]         # Allowed CORS origins (default: all)
+  enable_swagger: false          # Enable Swagger UI for API docs
+
+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)
+  llama_executable: llama-server # Path to llama-server executable
+  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
+```
+
+<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
 
@@ -128,16 +178,19 @@ server:
 
 ```yaml
 instances:
-  port_range: [8000, 9000]           # Port range for instances (default: [8000, 9000])
+  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)
+  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)
+  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)
+  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)
+  auto_create_dirs: true                            # Automatically create data/config/logs directories (default: true)
-  max_instances: -1                  # Maximum instances (-1 = unlimited)
+  max_instances: -1                                 # Maximum instances (-1 = unlimited)
-  llama_executable: "llama-server"   # Path to llama-server executable
+  llama_executable: "llama-server"                  # Path to llama-server executable
-  default_auto_restart: true         # Default auto-restart setting
+  default_auto_restart: true                        # Default auto-restart setting
-  default_max_restarts: 3            # Default maximum restart attempts
+  default_max_restarts: 3                           # Default maximum restart attempts
-  default_restart_delay: 5           # Default restart delay in seconds
+  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:**
@@ -151,6 +204,10 @@ 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
 
@@ -168,147 +225,8 @@ auth:
 - `LLAMACTL_REQUIRE_MANAGEMENT_AUTH` - Require auth for management endpoints (true/false)
 - `LLAMACTL_MANAGEMENT_KEYS` - Comma-separated management API keys
 
-### Example Configuration
+</details>
-
-```yaml
-server:
-  host: "0.0.0.0"
-  port: 8080
-
-instances:
-  port_range: [8001, 8100]
-  data_dir: "/var/lib/llamactl"
-  configs_dir: "/var/lib/llamactl/instances"
-  logs_dir: "/var/log/llamactl"
-  auto_create_dirs: true
-  max_instances: 10
-  llama_executable: "/usr/local/bin/llama-server"
-  default_auto_restart: true
-  default_max_restarts: 5
-  default_restart_delay: 10
-
-auth:
-  require_inference_auth: true
-  inference_keys: ["sk-inference-abc123"]
-  require_management_auth: true
-  management_keys: ["sk-management-xyz456"]
-```
-
-## Usage
-
-### Starting the Server
-
-```bash
-# Start with default configuration
-./llamactl
-
-# Start with custom config file
-LLAMACTL_CONFIG_PATH=/path/to/config.yaml ./llamactl
-
-# Start with environment variables
-LLAMACTL_PORT=9090 LLAMACTL_LOG_DIR=/custom/logs ./llamactl
-```
-
-### Authentication
-
-llamactl supports API Key authentication for both management and inference (OpenAI-compatible) endpoints. There are separate keys for management and inference APIs:
-
-- **Management keys** grant full access to instance management
-- **Inference keys** grant access to OpenAI-compatible endpoints
-- Management keys also work for inference endpoints (higher privilege)
-
-**How to Use:**
-Pass your API key in requests using one of:
-- `Authorization: Bearer <key>` header
-- `X-API-Key: <key>` header
-- `api_key=<key>` query parameter
-
-**Auto-generated keys**: If no keys are set and authentication is required, a key will be generated and printed to the terminal at startup. For production, set your own keys in config or environment variables.
-
-### Web Dashboard
-
-Open your browser and navigate to `http://localhost:8080` to access the web dashboard.
-
-### API Usage
-
-The REST API is available at `http://localhost:8080/api/v1`. See the Swagger documentation at `http://localhost:8080/swagger/` for complete API reference.
-
-#### Create an Instance
-
-```bash
-curl -X POST http://localhost:8080/api/v1/instances/my-instance \
-  -H "Content-Type: application/json" \
-  -H "Authorization: Bearer sk-management-your-key" \
-  -d '{
-    "model": "/path/to/model.gguf",
-    "gpu_layers": 32,
-    "auto_restart": true
-  }'
-```
-
-#### List Instances
-
-```bash
-curl -H "Authorization: Bearer sk-management-your-key" \
-  http://localhost:8080/api/v1/instances
-```
-
-#### Start/Stop Instance
-
-```bash
-# Start
-curl -X POST \
-  -H "Authorization: Bearer sk-management-your-key" \
-  http://localhost:8080/api/v1/instances/my-instance/start
-
-# Stop
-curl -X POST \
-  -H "Authorization: Bearer sk-management-your-key" \
-  http://localhost:8080/api/v1/instances/my-instance/stop
-```
-
-### OpenAI Compatible Endpoints
-
-Route requests to instances by including the instance name as the model parameter:
-
-```bash
-curl -X POST http://localhost:8080/v1/chat/completions \
-  -H "Content-Type: application/json" \
-  -H "Authorization: Bearer sk-inference-your-key" \
-  -d '{
-    "model": "my-instance",
-    "messages": [{"role": "user", "content": "Hello!"}]
-  }'
-```
-
-## Development
-
-### Running Tests
-
-```bash
-# Go tests
-go test ./...
-
-# Web UI tests
-cd webui
-npm test
-```
-
-### Development Server
-
-```bash
-# Start Go server in development mode
-go run ./cmd/server
-
-# Start web UI development server (in another terminal)
-cd webui
-npm run dev
-```
-
-## API Documentation
-
-Interactive API documentation is available at `http://localhost:8080/swagger/` when the server is running.
 
 ## License
 
-This project is licensed under the MIT License. See the [LICENSE](LICENSE) file for details.
+MIT License - see [LICENSE](LICENSE) file.

apidocs/docs.go

@@ -1,5 +1,5 @@
-// Package docs Code generated by swaggo/swag. DO NOT EDIT
+// Package apidocs Code generated by swaggo/swag. DO NOT EDIT
-package docs
+package apidocs
 
 import "github.com/swaggo/swag"
 
@@ -37,7 +37,7 @@ const docTemplate = `{
                         "schema": {
                             "type": "array",
                             "items": {
-                                "$ref": "#/definitions/llamactl.Instance"
+                                "$ref": "#/definitions/instance.Process"
                             }
                         }
                     },
@@ -75,7 +75,7 @@ const docTemplate = `{
                     "200": {
                         "description": "Instance details",
                         "schema": {
-                            "$ref": "#/definitions/llamactl.Instance"
+                            "$ref": "#/definitions/instance.Process"
                         }
                     },
                     "400": {
@@ -120,7 +120,7 @@ const docTemplate = `{
                         "in": "body",
                         "required": true,
                         "schema": {
-                            "$ref": "#/definitions/llamactl.CreateInstanceOptions"
+                            "$ref": "#/definitions/instance.CreateInstanceOptions"
                         }
                     }
                 ],
@@ -128,7 +128,7 @@ const docTemplate = `{
                     "200": {
                         "description": "Updated instance details",
                         "schema": {
-                            "$ref": "#/definitions/llamactl.Instance"
+                            "$ref": "#/definitions/instance.Process"
                         }
                     },
                     "400": {
@@ -173,7 +173,7 @@ const docTemplate = `{
                         "in": "body",
                         "required": true,
                         "schema": {
-                            "$ref": "#/definitions/llamactl.CreateInstanceOptions"
+                            "$ref": "#/definitions/instance.CreateInstanceOptions"
                         }
                     }
                 ],
@@ -181,7 +181,7 @@ const docTemplate = `{
                     "201": {
                         "description": "Created instance details",
                         "schema": {
-                            "$ref": "#/definitions/llamactl.Instance"
+                            "$ref": "#/definitions/instance.Process"
                         }
                     },
                     "400": {
@@ -401,7 +401,7 @@ const docTemplate = `{
                     "200": {
                         "description": "Restarted instance details",
                         "schema": {
-                            "$ref": "#/definitions/llamactl.Instance"
+                            "$ref": "#/definitions/instance.Process"
                         }
                     },
                     "400": {
@@ -444,7 +444,7 @@ const docTemplate = `{
                     "200": {
                         "description": "Started instance details",
                         "schema": {
-                            "$ref": "#/definitions/llamactl.Instance"
+                            "$ref": "#/definitions/instance.Process"
                         }
                     },
                     "400": {
@@ -487,7 +487,7 @@ const docTemplate = `{
                     "200": {
                         "description": "Stopped instance details",
                         "schema": {
-                            "$ref": "#/definitions/llamactl.Instance"
+                            "$ref": "#/definitions/instance.Process"
                         }
                     },
                     "400": {
@@ -639,7 +639,35 @@ const docTemplate = `{
                     "200": {
                         "description": "List of OpenAI-compatible instances",
                         "schema": {
-                            "$ref": "#/definitions/llamactl.OpenAIListInstancesResponse"
+                            "$ref": "#/definitions/server.OpenAIListInstancesResponse"
+                        }
+                    },
+                    "500": {
+                        "description": "Internal Server Error",
+                        "schema": {
+                            "type": "string"
+                        }
+                    }
+                }
+            }
+        },
+        "/version": {
+            "get": {
+                "security": [
+                    {
+                        "ApiKeyAuth": []
+                    }
+                ],
+                "description": "Returns the version of the llamactl command",
+                "tags": [
+                    "version"
+                ],
+                "summary": "Get llamactl version",
+                "responses": {
+                    "200": {
+                        "description": "Version information",
+                        "schema": {
+                            "type": "string"
                         }
                     },
                     "500": {
@@ -653,7 +681,7 @@ const docTemplate = `{
         }
     },
     "definitions": {
-        "llamactl.CreateInstanceOptions": {
+        "instance.CreateInstanceOptions": {
             "type": "object",
             "properties": {
                 "alias": {
@@ -751,7 +779,6 @@ const docTemplate = `{
                     "type": "string"
                 },
                 "draft_max": {
-                    "description": "Speculative decoding params",
                     "type": "integer"
                 },
                 "draft_min": {
@@ -955,7 +982,7 @@ const docTemplate = `{
                     "type": "boolean"
                 },
                 "no_context_shift": {
-                    "description": "Server/Example-specific params",
+                    "description": "Example-specific params",
                     "type": "boolean"
                 },
                 "no_escape": {
@@ -1027,10 +1054,10 @@ const docTemplate = `{
                 "presence_penalty": {
                     "type": "number"
                 },
-                "priority": {
+                "prio": {
                     "type": "integer"
                 },
-                "priority_batch": {
+                "prio_batch": {
                     "type": "integer"
                 },
                 "props": {
@@ -1101,7 +1128,7 @@ const docTemplate = `{
                 "ssl_key_file": {
                     "type": "string"
                 },
-                "temperature": {
+                "temp": {
                     "type": "number"
                 },
                 "tensor_split": {
@@ -1167,7 +1194,7 @@ const docTemplate = `{
                 }
             }
         },
-        "llamactl.Instance": {
+        "instance.Process": {
             "type": "object",
             "properties": {
                 "created": {
@@ -1183,7 +1210,7 @@ const docTemplate = `{
                 }
             }
         },
-        "llamactl.OpenAIInstance": {
+        "server.OpenAIInstance": {
             "type": "object",
             "properties": {
                 "created": {
@@ -1200,13 +1227,13 @@ const docTemplate = `{
                 }
             }
         },
-        "llamactl.OpenAIListInstancesResponse": {
+        "server.OpenAIListInstancesResponse": {
             "type": "object",
             "properties": {
                 "data": {
                     "type": "array",
                     "items": {
-                        "$ref": "#/definitions/llamactl.OpenAIInstance"
+                        "$ref": "#/definitions/server.OpenAIInstance"
                     }
                 },
                 "object": {

apidocs/swagger.json

@@ -30,7 +30,7 @@
                         "schema": {
                             "type": "array",
                             "items": {
-                                "$ref": "#/definitions/llamactl.Instance"
+                                "$ref": "#/definitions/instance.Process"
                             }
                         }
                     },
@@ -68,7 +68,7 @@
                     "200": {
                         "description": "Instance details",
                         "schema": {
-                            "$ref": "#/definitions/llamactl.Instance"
+                            "$ref": "#/definitions/instance.Process"
                         }
                     },
                     "400": {
@@ -113,7 +113,7 @@
                         "in": "body",
                         "required": true,
                         "schema": {
-                            "$ref": "#/definitions/llamactl.CreateInstanceOptions"
+                            "$ref": "#/definitions/instance.CreateInstanceOptions"
                         }
                     }
                 ],
@@ -121,7 +121,7 @@
                     "200": {
                         "description": "Updated instance details",
                         "schema": {
-                            "$ref": "#/definitions/llamactl.Instance"
+                            "$ref": "#/definitions/instance.Process"
                         }
                     },
                     "400": {
@@ -166,7 +166,7 @@
                         "in": "body",
                         "required": true,
                         "schema": {
-                            "$ref": "#/definitions/llamactl.CreateInstanceOptions"
+                            "$ref": "#/definitions/instance.CreateInstanceOptions"
                         }
                     }
                 ],
@@ -174,7 +174,7 @@
                     "201": {
                         "description": "Created instance details",
                         "schema": {
-                            "$ref": "#/definitions/llamactl.Instance"
+                            "$ref": "#/definitions/instance.Process"
                         }
                     },
                     "400": {
@@ -394,7 +394,7 @@
                     "200": {
                         "description": "Restarted instance details",
                         "schema": {
-                            "$ref": "#/definitions/llamactl.Instance"
+                            "$ref": "#/definitions/instance.Process"
                         }
                     },
                     "400": {
@@ -437,7 +437,7 @@
                     "200": {
                         "description": "Started instance details",
                         "schema": {
-                            "$ref": "#/definitions/llamactl.Instance"
+                            "$ref": "#/definitions/instance.Process"
                         }
                     },
                     "400": {
@@ -480,7 +480,7 @@
                     "200": {
                         "description": "Stopped instance details",
                         "schema": {
-                            "$ref": "#/definitions/llamactl.Instance"
+                            "$ref": "#/definitions/instance.Process"
                         }
                     },
                     "400": {
@@ -632,7 +632,35 @@
                     "200": {
                         "description": "List of OpenAI-compatible instances",
                         "schema": {
-                            "$ref": "#/definitions/llamactl.OpenAIListInstancesResponse"
+                            "$ref": "#/definitions/server.OpenAIListInstancesResponse"
+                        }
+                    },
+                    "500": {
+                        "description": "Internal Server Error",
+                        "schema": {
+                            "type": "string"
+                        }
+                    }
+                }
+            }
+        },
+        "/version": {
+            "get": {
+                "security": [
+                    {
+                        "ApiKeyAuth": []
+                    }
+                ],
+                "description": "Returns the version of the llamactl command",
+                "tags": [
+                    "version"
+                ],
+                "summary": "Get llamactl version",
+                "responses": {
+                    "200": {
+                        "description": "Version information",
+                        "schema": {
+                            "type": "string"
                         }
                     },
                     "500": {
@@ -646,7 +674,7 @@
         }
     },
     "definitions": {
-        "llamactl.CreateInstanceOptions": {
+        "instance.CreateInstanceOptions": {
             "type": "object",
             "properties": {
                 "alias": {
@@ -744,7 +772,6 @@
                     "type": "string"
                 },
                 "draft_max": {
-                    "description": "Speculative decoding params",
                     "type": "integer"
                 },
                 "draft_min": {
@@ -948,7 +975,7 @@
                     "type": "boolean"
                 },
                 "no_context_shift": {
-                    "description": "Server/Example-specific params",
+                    "description": "Example-specific params",
                     "type": "boolean"
                 },
                 "no_escape": {
@@ -1020,10 +1047,10 @@
                 "presence_penalty": {
                     "type": "number"
                 },
-                "priority": {
+                "prio": {
                     "type": "integer"
                 },
-                "priority_batch": {
+                "prio_batch": {
                     "type": "integer"
                 },
                 "props": {
@@ -1094,7 +1121,7 @@
                 "ssl_key_file": {
                     "type": "string"
                 },
-                "temperature": {
+                "temp": {
                     "type": "number"
                 },
                 "tensor_split": {
@@ -1160,7 +1187,7 @@
                 }
             }
         },
-        "llamactl.Instance": {
+        "instance.Process": {
             "type": "object",
             "properties": {
                 "created": {
@@ -1176,7 +1203,7 @@
                 }
             }
         },
-        "llamactl.OpenAIInstance": {
+        "server.OpenAIInstance": {
             "type": "object",
             "properties": {
                 "created": {
@@ -1193,13 +1220,13 @@
                 }
             }
         },
-        "llamactl.OpenAIListInstancesResponse": {
+        "server.OpenAIListInstancesResponse": {
             "type": "object",
             "properties": {
                 "data": {
                     "type": "array",
                     "items": {
-                        "$ref": "#/definitions/llamactl.OpenAIInstance"
+                        "$ref": "#/definitions/server.OpenAIInstance"
                     }
                 },
                 "object": {

apidocs/swagger.yaml

@@ -1,6 +1,6 @@
 basePath: /api/v1
 definitions:
-  llamactl.CreateInstanceOptions:
+  instance.CreateInstanceOptions:
     properties:
       alias:
         type: string
@@ -66,7 +66,6 @@ definitions:
       device_draft:
         type: string
       draft_max:
-        description: Speculative decoding params
         type: integer
       draft_min:
         type: integer
@@ -203,7 +202,7 @@ definitions:
       no_cont_batching:
         type: boolean
       no_context_shift:
-        description: Server/Example-specific params
+        description: Example-specific params
         type: boolean
       no_escape:
         type: boolean
@@ -251,9 +250,9 @@ definitions:
         type: integer
       presence_penalty:
         type: number
-      priority:
+      prio:
         type: integer
-      priority_batch:
+      prio_batch:
         type: integer
       props:
         type: boolean
@@ -301,7 +300,7 @@ definitions:
         type: string
       ssl_key_file:
         type: string
-      temperature:
+      temp:
         type: number
       tensor_split:
         type: string
@@ -345,7 +344,7 @@ definitions:
       yarn_orig_ctx:
         type: integer
     type: object
-  llamactl.Instance:
+  instance.Process:
     properties:
       created:
         description: Creation time
@@ -356,7 +355,7 @@ definitions:
         description: Status
         type: boolean
     type: object
-  llamactl.OpenAIInstance:
+  server.OpenAIInstance:
     properties:
       created:
         type: integer
@@ -367,11 +366,11 @@ definitions:
       owned_by:
         type: string
     type: object
-  llamactl.OpenAIListInstancesResponse:
+  server.OpenAIListInstancesResponse:
     properties:
       data:
         items:
-          $ref: '#/definitions/llamactl.OpenAIInstance'
+          $ref: '#/definitions/server.OpenAIInstance'
         type: array
       object:
         type: string
@@ -393,7 +392,7 @@ paths:
           description: List of instances
           schema:
             items:
-              $ref: '#/definitions/llamactl.Instance'
+              $ref: '#/definitions/instance.Process'
             type: array
         "500":
           description: Internal Server Error
@@ -441,7 +440,7 @@ paths:
         "200":
           description: Instance details
           schema:
-            $ref: '#/definitions/llamactl.Instance'
+            $ref: '#/definitions/instance.Process'
         "400":
           description: Invalid name format
           schema:
@@ -470,12 +469,12 @@ paths:
         name: options
         required: true
         schema:
-          $ref: '#/definitions/llamactl.CreateInstanceOptions'
+          $ref: '#/definitions/instance.CreateInstanceOptions'
       responses:
         "201":
           description: Created instance details
           schema:
-            $ref: '#/definitions/llamactl.Instance'
+            $ref: '#/definitions/instance.Process'
         "400":
           description: Invalid request body
           schema:
@@ -504,12 +503,12 @@ paths:
         name: options
         required: true
         schema:
-          $ref: '#/definitions/llamactl.CreateInstanceOptions'
+          $ref: '#/definitions/instance.CreateInstanceOptions'
       responses:
         "200":
           description: Updated instance details
           schema:
-            $ref: '#/definitions/llamactl.Instance'
+            $ref: '#/definitions/instance.Process'
         "400":
           description: Invalid name format
           schema:
@@ -627,7 +626,7 @@ paths:
         "200":
           description: Restarted instance details
           schema:
-            $ref: '#/definitions/llamactl.Instance'
+            $ref: '#/definitions/instance.Process'
         "400":
           description: Invalid name format
           schema:
@@ -654,7 +653,7 @@ paths:
         "200":
           description: Started instance details
           schema:
-            $ref: '#/definitions/llamactl.Instance'
+            $ref: '#/definitions/instance.Process'
         "400":
           description: Invalid name format
           schema:
@@ -681,7 +680,7 @@ paths:
         "200":
           description: Stopped instance details
           schema:
-            $ref: '#/definitions/llamactl.Instance'
+            $ref: '#/definitions/instance.Process'
         "400":
           description: Invalid name format
           schema:
@@ -777,7 +776,7 @@ paths:
         "200":
           description: List of OpenAI-compatible instances
           schema:
-            $ref: '#/definitions/llamactl.OpenAIListInstancesResponse'
+            $ref: '#/definitions/server.OpenAIListInstancesResponse'
         "500":
           description: Internal Server Error
           schema:
@@ -787,4 +786,21 @@ paths:
       summary: List instances in OpenAI-compatible format
       tags:
       - openai
+  /version:
+    get:
+      description: Returns the version of the llamactl command
+      responses:
+        "200":
+          description: Version information
+          schema:
+            type: string
+        "500":
+          description: Internal Server Error
+          schema:
+            type: string
+      security:
+      - ApiKeyAuth: []
+      summary: Get llamactl version
+      tags:
+      - version
 swagger: "2.0"

cmd/server/main.go

@@ -11,6 +11,11 @@ import (
 	"syscall"
 )
 
+// version is set at build time using -ldflags "-X main.version=1.0.0"
+var version string = "unknown"
+var commitHash string = "unknown"
+var buildTime string = "unknown"
+
 // @title llamactl API
 // @version 1.0
 // @description llamactl is a control server for managing Llama Server instances.
@@ -19,6 +24,14 @@ import (
 // @basePath /api/v1
 func main() {
 
+	// --version flag to print the version
+	if len(os.Args) > 1 && os.Args[1] == "--version" {
+		fmt.Printf("llamactl version: %s\n", version)
+		fmt.Printf("Commit hash: %s\n", commitHash)
+		fmt.Printf("Build time: %s\n", buildTime)
+		return
+	}
+
 	configPath := os.Getenv("LLAMACTL_CONFIG_PATH")
 	cfg, err := config.LoadConfig(configPath)
 	if err != nil {
@@ -26,6 +39,11 @@ func main() {
 		fmt.Println("Using default configuration.")
 	}
 
+	// Set version information
+	cfg.Version = version
+	cfg.CommitHash = commitHash
+	cfg.BuildTime = buildTime
+
 	// Create the data directory if it doesn't exist
 	if cfg.Instances.AutoCreateDirs {
 		if err := os.MkdirAll(cfg.Instances.InstancesDir, 0755); err != nil {

pkg/backends/llamacpp/llama.go

@@ -15,12 +15,12 @@ type LlamaServerOptions struct {
 	CPUMask                 string   `json:"cpu_mask,omitempty"`
 	CPURange                string   `json:"cpu_range,omitempty"`
 	CPUStrict               int      `json:"cpu_strict,omitempty"`
-	Priority                int      `json:"priority,omitempty"`
+	Prio                    int      `json:"prio,omitempty"`
 	Poll                    int      `json:"poll,omitempty"`
 	CPUMaskBatch            string   `json:"cpu_mask_batch,omitempty"`
 	CPURangeBatch           string   `json:"cpu_range_batch,omitempty"`
 	CPUStrictBatch          int      `json:"cpu_strict_batch,omitempty"`
-	PriorityBatch           int      `json:"priority_batch,omitempty"`
+	PrioBatch               int      `json:"prio_batch,omitempty"`
 	PollBatch               int      `json:"poll_batch,omitempty"`
 	CtxSize                 int      `json:"ctx_size,omitempty"`
 	Predict                 int      `json:"predict,omitempty"`
@@ -83,7 +83,7 @@ type LlamaServerOptions struct {
 	Seed               int      `json:"seed,omitempty"`
 	SamplingSeq        string   `json:"sampling_seq,omitempty"`
 	IgnoreEOS          bool     `json:"ignore_eos,omitempty"`
-	Temperature        float64  `json:"temperature,omitempty"`
+	Temperature        float64  `json:"temp,omitempty"`
 	TopK               int      `json:"top_k,omitempty"`
 	TopP               float64  `json:"top_p,omitempty"`
 	MinP               float64  `json:"min_p,omitempty"`
@@ -110,7 +110,7 @@ type LlamaServerOptions struct {
 	JSONSchema         string   `json:"json_schema,omitempty"`
 	JSONSchemaFile     string   `json:"json_schema_file,omitempty"`
 
-	// Server/Example-specific params
+	// Example-specific params
 	NoContextShift       bool    `json:"no_context_shift,omitempty"`
 	Special              bool    `json:"special,omitempty"`
 	NoWarmup             bool    `json:"no_warmup,omitempty"`
@@ -150,17 +150,15 @@ type LlamaServerOptions struct {
 	NoPrefillAssistant   bool    `json:"no_prefill_assistant,omitempty"`
 	SlotPromptSimilarity float64 `json:"slot_prompt_similarity,omitempty"`
 	LoraInitWithoutApply bool    `json:"lora_init_without_apply,omitempty"`
-
+	DraftMax             int     `json:"draft_max,omitempty"`
-	// Speculative decoding params
+	DraftMin             int     `json:"draft_min,omitempty"`
-	DraftMax        int     `json:"draft_max,omitempty"`
+	DraftPMin            float64 `json:"draft_p_min,omitempty"`
-	DraftMin        int     `json:"draft_min,omitempty"`
+	CtxSizeDraft         int     `json:"ctx_size_draft,omitempty"`
-	DraftPMin       float64 `json:"draft_p_min,omitempty"`
+	DeviceDraft          string  `json:"device_draft,omitempty"`
-	CtxSizeDraft    int     `json:"ctx_size_draft,omitempty"`
+	GPULayersDraft       int     `json:"gpu_layers_draft,omitempty"`
-	DeviceDraft     string  `json:"device_draft,omitempty"`
+	ModelDraft           string  `json:"model_draft,omitempty"`
-	GPULayersDraft  int     `json:"gpu_layers_draft,omitempty"`
+	CacheTypeKDraft      string  `json:"cache_type_k_draft,omitempty"`
-	ModelDraft      string  `json:"model_draft,omitempty"`
+	CacheTypeVDraft      string  `json:"cache_type_v_draft,omitempty"`
-	CacheTypeKDraft string  `json:"cache_type_k_draft,omitempty"`
-	CacheTypeVDraft string  `json:"cache_type_v_draft,omitempty"`
 
 	// Audio/TTS params
 	ModelVocoder      string `json:"model_vocoder,omitempty"`
@@ -199,62 +197,75 @@ func (o *LlamaServerOptions) UnmarshalJSON(data []byte) error {
 
 	// Handle alternative field names
 	fieldMappings := map[string]string{
-		// Official llama-server short forms from the documentation
+		// Common params
-		"t":    "threads",                // -t, --threads N
+		"t":             "threads",         // -t, --threads N
-		"tb":   "threads_batch",          // -tb, --threads-batch N
+		"tb":            "threads_batch",   // -tb, --threads-batch N
-		"C":    "cpu_mask",               // -C, --cpu-mask M
+		"C":             "cpu_mask",        // -C, --cpu-mask M
-		"Cr":   "cpu_range",              // -Cr, --cpu-range lo-hi
+		"Cr":            "cpu_range",       // -Cr, --cpu-range lo-hi
-		"Cb":   "cpu_mask_batch",         // -Cb, --cpu-mask-batch M
+		"Cb":            "cpu_mask_batch",  // -Cb, --cpu-mask-batch M
-		"Crb":  "cpu_range_batch",        // -Crb, --cpu-range-batch lo-hi
+		"Crb":           "cpu_range_batch", // -Crb, --cpu-range-batch lo-hi
-		"c":    "ctx_size",               // -c, --ctx-size N
+		"c":             "ctx_size",        // -c, --ctx-size N
-		"n":    "predict",                // -n, --predict, --n-predict N
+		"n":             "predict",         // -n, --predict N
-		"b":    "batch_size",             // -b, --batch-size N
+		"n-predict":     "predict",         // --n-predict N
-		"ub":   "ubatch_size",            // -ub, --ubatch-size N
+		"b":             "batch_size",      // -b, --batch-size N
-		"fa":   "flash_attn",             // -fa, --flash-attn
+		"ub":            "ubatch_size",     // -ub, --ubatch-size N
-		"e":    "escape",                 // -e, --escape
+		"fa":            "flash_attn",      // -fa, --flash-attn
-		"dkvc": "dump_kv_cache",          // -dkvc, --dump-kv-cache
+		"e":             "escape",          // -e, --escape
-		"nkvo": "no_kv_offload",          // -nkvo, --no-kv-offload
+		"dkvc":          "dump_kv_cache",   // -dkvc, --dump-kv-cache
-		"ctk":  "cache_type_k",           // -ctk, --cache-type-k TYPE
+		"nkvo":          "no_kv_offload",   // -nkvo, --no-kv-offload
-		"ctv":  "cache_type_v",           // -ctv, --cache-type-v TYPE
+		"ctk":           "cache_type_k",    // -ctk, --cache-type-k TYPE
-		"dt":   "defrag_thold",           // -dt, --defrag-thold N
+		"ctv":           "cache_type_v",    // -ctv, --cache-type-v TYPE
-		"np":   "parallel",               // -np, --parallel N
+		"dt":            "defrag_thold",    // -dt, --defrag-thold N
-		"dev":  "device",                 // -dev, --device <dev1,dev2,..>
+		"np":            "parallel",        // -np, --parallel N
-		"ot":   "override_tensor",        // --override-tensor, -ot
+		"dev":           "device",          // -dev, --device <dev1,dev2,..>
-		"ngl":  "gpu_layers",             // -ngl, --gpu-layers, --n-gpu-layers N
+		"ot":            "override_tensor", // --override-tensor, -ot
-		"sm":   "split_mode",             // -sm, --split-mode
+		"ngl":           "gpu_layers",      // -ngl, --gpu-layers, --n-gpu-layers N
-		"ts":   "tensor_split",           // -ts, --tensor-split N0,N1,N2,...
+		"n-gpu-layers":  "gpu_layers",      // --n-gpu-layers N
-		"mg":   "main_gpu",               // -mg, --main-gpu INDEX
+		"sm":            "split_mode",      // -sm, --split-mode
-		"m":    "model",                  // -m, --model FNAME
+		"ts":            "tensor_split",    // -ts, --tensor-split N0,N1,N2,...
-		"mu":   "model_url",              // -mu, --model-url MODEL_URL
+		"mg":            "main_gpu",        // -mg, --main-gpu INDEX
-		"hf":   "hf_repo",                // -hf, -hfr, --hf-repo
+		"m":             "model",           // -m, --model FNAME
-		"hfr":  "hf_repo",                // -hf, -hfr, --hf-repo
+		"mu":            "model_url",       // -mu, --model-url MODEL_URL
-		"hfd":  "hf_repo_draft",          // -hfd, -hfrd, --hf-repo-draft
+		"hf":            "hf_repo",         // -hf, -hfr, --hf-repo
-		"hfrd": "hf_repo_draft",          // -hfd, -hfrd, --hf-repo-draft
+		"hfr":           "hf_repo",         // -hf, -hfr, --hf-repo
-		"hff":  "hf_file",                // -hff, --hf-file FILE
+		"hfd":           "hf_repo_draft",   // -hfd, -hfrd, --hf-repo-draft
-		"hfv":  "hf_repo_v",              // -hfv, -hfrv, --hf-repo-v
+		"hfrd":          "hf_repo_draft",   // -hfd, -hfrd, --hf-repo-draft
-		"hfrv": "hf_repo_v",              // -hfv, -hfrv, --hf-repo-v
+		"hff":           "hf_file",         // -hff, --hf-file FILE
-		"hffv": "hf_file_v",              // -hffv, --hf-file-v FILE
+		"hfv":           "hf_repo_v",       // -hfv, -hfrv, --hf-repo-v
-		"hft":  "hf_token",               // -hft, --hf-token TOKEN
+		"hfrv":          "hf_repo_v",       // -hfv, -hfrv, --hf-repo-v
-		"v":    "verbose",                // -v, --verbose, --log-verbose
+		"hffv":          "hf_file_v",       // -hffv, --hf-file-v FILE
-		"lv":   "verbosity",              // -lv, --verbosity, --log-verbosity N
+		"hft":           "hf_token",        // -hft, --hf-token TOKEN
-		"s":    "seed",                   // -s, --seed SEED
+		"v":             "verbose",         // -v, --verbose, --log-verbose
-		"temp": "temperature",            // --temp N
+		"log-verbose":   "verbose",         // --log-verbose
-		"l":    "logit_bias",             // -l, --logit-bias
+		"lv":            "verbosity",       // -lv, --verbosity, --log-verbosity N
-		"j":    "json_schema",            // -j, --json-schema SCHEMA
+		"log-verbosity": "verbosity",       // --log-verbosity N
-		"jf":   "json_schema_file",       // -jf, --json-schema-file FILE
+
-		"sp":   "special",                // -sp, --special
+		// Sampling params
-		"cb":   "cont_batching",          // -cb, --cont-batching
+		"s":  "seed",             // -s, --seed SEED
-		"nocb": "no_cont_batching",       // -nocb, --no-cont-batching
+		"l":  "logit_bias",       // -l, --logit-bias
-		"a":    "alias",                  // -a, --alias STRING
+		"j":  "json_schema",      // -j, --json-schema SCHEMA
-		"to":   "timeout",                // -to, --timeout N
+		"jf": "json_schema_file", // -jf, --json-schema-file FILE
-		"sps":  "slot_prompt_similarity", // -sps, --slot-prompt-similarity
+
-		"cd":   "ctx_size_draft",         // -cd, --ctx-size-draft N
+		// Example-specific params
-		"devd": "device_draft",           // -devd, --device-draft
+		"sp":                 "special",                // -sp, --special
-		"ngld": "gpu_layers_draft",       // -ngld, --gpu-layers-draft
+		"cb":                 "cont_batching",          // -cb, --cont-batching
-		"md":   "model_draft",            // -md, --model-draft FNAME
+		"nocb":               "no_cont_batching",       // -nocb, --no-cont-batching
-		"ctkd": "cache_type_k_draft",     // -ctkd, --cache-type-k-draft TYPE
+		"a":                  "alias",                  // -a, --alias STRING
-		"ctvd": "cache_type_v_draft",     // -ctvd, --cache-type-v-draft TYPE
+		"embeddings":         "embedding",              // --embeddings
-		"mv":   "model_vocoder",          // -mv, --model-vocoder FNAME
+		"rerank":             "reranking",              // --reranking
+		"to":                 "timeout",                // -to, --timeout N
+		"sps":                "slot_prompt_similarity", // -sps, --slot-prompt-similarity
+		"draft":              "draft-max",              // -draft, --draft-max N
+		"draft-n":            "draft-max",              // --draft-n-max N
+		"draft-n-min":        "draft_min",              // --draft-n-min N
+		"cd":                 "ctx_size_draft",         // -cd, --ctx-size-draft N
+		"devd":               "device_draft",           // -devd, --device-draft
+		"ngld":               "gpu_layers_draft",       // -ngld, --gpu-layers-draft
+		"n-gpu-layers-draft": "gpu_layers_draft",       // --n-gpu-layers-draft N
+		"md":                 "model_draft",            // -md, --model-draft FNAME
+		"ctkd":               "cache_type_k_draft",     // -ctkd, --cache-type-k-draft TYPE
+		"ctvd":               "cache_type_v_draft",     // -ctvd, --cache-type-v-draft TYPE
+		"mv":                 "model_vocoder",          // -mv, --model-vocoder FNAME
 	}
 
 	// Process alternative field names

pkg/backends/llamacpp/llama_test.go

@@ -109,13 +109,13 @@ func TestBuildCommandArgs_NumericFields(t *testing.T) {
 	args := options.BuildCommandArgs()
 
 	expectedPairs := map[string]string{
-		"--port":        "8080",
+		"--port":       "8080",
-		"--threads":     "4",
+		"--threads":    "4",
-		"--ctx-size":    "2048",
+		"--ctx-size":   "2048",
-		"--gpu-layers":  "16",
+		"--gpu-layers": "16",
-		"--temperature": "0.7",
+		"--temp":       "0.7",
-		"--top-k":       "40",
+		"--top-k":      "40",
-		"--top-p":       "0.9",
+		"--top-p":      "0.9",
 	}
 
 	for flag, expectedValue := range expectedPairs {
@@ -231,7 +231,7 @@ func TestUnmarshalJSON_StandardFields(t *testing.T) {
 		"verbose": true,
 		"ctx_size": 4096,
 		"gpu_layers": 32,
-		"temperature": 0.7
+		"temp": 0.7
 	}`
 
 	var options llamacpp.LlamaServerOptions

pkg/config/config.go

@@ -12,9 +12,12 @@ import (
 
 // AppConfig represents the configuration for llamactl
 type AppConfig struct {
-	Server    ServerConfig    `yaml:"server"`
+	Server     ServerConfig    `yaml:"server"`
-	Instances InstancesConfig `yaml:"instances"`
+	Instances  InstancesConfig `yaml:"instances"`
-	Auth      AuthConfig      `yaml:"auth"`
+	Auth       AuthConfig      `yaml:"auth"`
+	Version    string          `yaml:"-"`
+	CommitHash string          `yaml:"-"`
+	BuildTime  string          `yaml:"-"`
 }
 
 // ServerConfig contains HTTP server configuration
@@ -63,6 +66,15 @@ type InstancesConfig struct {
 
 	// Default restart delay for new instances (in seconds)
 	DefaultRestartDelay int `yaml:"default_restart_delay"`
+
+	// Default on-demand start setting for new instances
+	DefaultOnDemandStart bool `yaml:"default_on_demand_start"`
+
+	// How long to wait for an instance to start on demand (in seconds)
+	OnDemandStartTimeout int `yaml:"on_demand_start_timeout,omitempty"`
+
+	// Interval for checking instance timeouts (in minutes)
+	TimeoutCheckInterval int `yaml:"timeout_check_interval"`
 }
 
 // AuthConfig contains authentication settings
@@ -95,16 +107,19 @@ func LoadConfig(configPath string) (AppConfig, error) {
 			EnableSwagger:  false,
 		},
 		Instances: InstancesConfig{
-			PortRange:           [2]int{8000, 9000},
+			PortRange:            [2]int{8000, 9000},
-			DataDir:             getDefaultDataDirectory(),
+			DataDir:              getDefaultDataDirectory(),
-			InstancesDir:        filepath.Join(getDefaultDataDirectory(), "instances"),
+			InstancesDir:         filepath.Join(getDefaultDataDirectory(), "instances"),
-			LogsDir:             filepath.Join(getDefaultDataDirectory(), "logs"),
+			LogsDir:              filepath.Join(getDefaultDataDirectory(), "logs"),
-			AutoCreateDirs:      true,
+			AutoCreateDirs:       true,
-			MaxInstances:        -1, // -1 means unlimited
+			MaxInstances:         -1, // -1 means unlimited
-			LlamaExecutable:     "llama-server",
+			LlamaExecutable:      "llama-server",
-			DefaultAutoRestart:  true,
+			DefaultAutoRestart:   true,
-			DefaultMaxRestarts:  3,
+			DefaultMaxRestarts:   3,
-			DefaultRestartDelay: 5,
+			DefaultRestartDelay:  5,
+			DefaultOnDemandStart: true,
+			OnDemandStartTimeout: 120, // 2 minutes
+			TimeoutCheckInterval: 5,   // Check timeouts every 5 minutes
 		},
 		Auth: AuthConfig{
 			RequireInferenceAuth:  true,
@@ -214,6 +229,21 @@ func loadEnvVars(cfg *AppConfig) {
 			cfg.Instances.DefaultRestartDelay = seconds
 		}
 	}
+	if onDemandStart := os.Getenv("LLAMACTL_DEFAULT_ON_DEMAND_START"); onDemandStart != "" {
+		if b, err := strconv.ParseBool(onDemandStart); err == nil {
+			cfg.Instances.DefaultOnDemandStart = b
+		}
+	}
+	if onDemandTimeout := os.Getenv("LLAMACTL_ON_DEMAND_START_TIMEOUT"); onDemandTimeout != "" {
+		if seconds, err := strconv.Atoi(onDemandTimeout); err == nil {
+			cfg.Instances.OnDemandStartTimeout = seconds
+		}
+	}
+	if timeoutCheckInterval := os.Getenv("LLAMACTL_TIMEOUT_CHECK_INTERVAL"); timeoutCheckInterval != "" {
+		if minutes, err := strconv.Atoi(timeoutCheckInterval); err == nil {
+			cfg.Instances.TimeoutCheckInterval = minutes
+		}
+	}
 	// Auth config
 	if requireInferenceAuth := os.Getenv("LLAMACTL_REQUIRE_INFERENCE_AUTH"); requireInferenceAuth != "" {
 		if b, err := strconv.ParseBool(requireInferenceAuth); err == nil {

pkg/instance/instance.go

@@ -13,16 +13,32 @@ import (
 	"net/url"
 	"os/exec"
 	"sync"
+	"sync/atomic"
 	"time"
 )
 
+// TimeProvider interface allows for testing with mock time
+type TimeProvider interface {
+	Now() time.Time
+}
+
+// realTimeProvider implements TimeProvider using the actual time
+type realTimeProvider struct{}
+
+func (realTimeProvider) Now() time.Time {
+	return time.Now()
+}
+
 type CreateInstanceOptions struct {
 	// Auto restart
-	AutoRestart *bool `json:"auto_restart,omitempty"`
+	AutoRestart  *bool `json:"auto_restart,omitempty"`
-	MaxRestarts *int  `json:"max_restarts,omitempty"`
+	MaxRestarts  *int  `json:"max_restarts,omitempty"`
-	// RestartDelay duration in seconds
+	RestartDelay *int  `json:"restart_delay,omitempty"`
-	RestartDelay *int `json:"restart_delay_seconds,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"`
 }
 
@@ -32,9 +48,11 @@ type CreateInstanceOptions struct {
 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"`
+		AutoRestart   *bool `json:"auto_restart,omitempty"`
-		MaxRestarts  *int  `json:"max_restarts,omitempty"`
+		MaxRestarts   *int  `json:"max_restarts,omitempty"`
-		RestartDelay *int  `json:"restart_delay_seconds,omitempty"`
+		RestartDelay  *int  `json:"restart_delay,omitempty"`
+		OnDemandStart *bool `json:"on_demand_start,omitempty"`
+		IdleTimeout   *int  `json:"idle_timeout,omitempty"`
 	}
 
 	var temp tempCreateOptions
@@ -46,6 +64,8 @@ func (c *CreateInstanceOptions) UnmarshalJSON(data []byte) error {
 	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 {
@@ -83,6 +103,10 @@ type Process struct {
 	// Restart control
 	restartCancel context.CancelFunc `json:"-"` // Cancel function for pending restarts
 	monitorDone   chan struct{}      `json:"-"` // Channel to signal monitor goroutine completion
+
+	// Timeout management
+	lastRequestTime atomic.Int64 // Unix timestamp of last request
+	timeProvider    TimeProvider `json:"-"` // Time provider for testing
 }
 
 // validateAndCopyOptions validates and creates a deep copy of the provided options
@@ -117,6 +141,20 @@ func validateAndCopyOptions(name string, options *CreateInstanceOptions) *Create
 			}
 			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
@@ -142,6 +180,16 @@ func applyDefaultOptions(options *CreateInstanceOptions, globalSettings *config.
 		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
@@ -158,10 +206,8 @@ func NewInstance(name string, globalSettings *config.InstancesConfig, options *C
 		options:        optionsCopy,
 		globalSettings: globalSettings,
 		logger:         logger,
-
+		timeProvider:   realTimeProvider{},
-		Running: false,
+		Created:        time.Now().Unix(),
-
-		Created: time.Now().Unix(),
 	}
 }
 
@@ -189,6 +235,11 @@ func (i *Process) SetOptions(options *CreateInstanceOptions) {
 	i.proxy = nil
 }
 
+// SetTimeProvider sets a custom time provider for testing
+func (i *Process) SetTimeProvider(tp TimeProvider) {
+	i.timeProvider = tp
+}
+
 // GetProxy returns the reverse proxy for this instance, creating it if needed
 func (i *Process) GetProxy() (*httputil.ReverseProxy, error) {
 	i.mu.Lock()

pkg/instance/instance_test.go

@@ -91,38 +91,6 @@ func TestNewInstance_WithRestartOptions(t *testing.T) {
 	}
 }
 
-func TestNewInstance_ValidationAndDefaults(t *testing.T) {
-	globalSettings := &config.InstancesConfig{
-		LogsDir:             "/tmp/test",
-		DefaultAutoRestart:  true,
-		DefaultMaxRestarts:  3,
-		DefaultRestartDelay: 5,
-	}
-
-	// Test with invalid negative values
-	invalidMaxRestarts := -5
-	invalidRestartDelay := -10
-
-	options := &instance.CreateInstanceOptions{
-		MaxRestarts:  &invalidMaxRestarts,
-		RestartDelay: &invalidRestartDelay,
-		LlamaServerOptions: llamacpp.LlamaServerOptions{
-			Model: "/path/to/model.gguf",
-		},
-	}
-
-	instance := instance.NewInstance("test-instance", globalSettings, options)
-	opts := instance.GetOptions()
-
-	// Check that negative values were corrected to 0
-	if opts.MaxRestarts == nil || *opts.MaxRestarts != 0 {
-		t.Errorf("Expected MaxRestarts to be corrected to 0, got %v", opts.MaxRestarts)
-	}
-	if opts.RestartDelay == nil || *opts.RestartDelay != 0 {
-		t.Errorf("Expected RestartDelay to be corrected to 0, got %v", opts.RestartDelay)
-	}
-}
-
 func TestSetOptions(t *testing.T) {
 	globalSettings := &config.InstancesConfig{
 		LogsDir:             "/tmp/test",
@@ -164,33 +132,6 @@ func TestSetOptions(t *testing.T) {
 	}
 }
 
-func TestSetOptions_NilOptions(t *testing.T) {
-	globalSettings := &config.InstancesConfig{
-		LogsDir:             "/tmp/test",
-		DefaultAutoRestart:  true,
-		DefaultMaxRestarts:  3,
-		DefaultRestartDelay: 5,
-	}
-
-	options := &instance.CreateInstanceOptions{
-		LlamaServerOptions: llamacpp.LlamaServerOptions{
-			Model: "/path/to/model.gguf",
-		},
-	}
-
-	instance := instance.NewInstance("test-instance", globalSettings, options)
-	originalOptions := instance.GetOptions()
-
-	// Try to set nil options
-	instance.SetOptions(nil)
-
-	// Options should remain unchanged
-	currentOptions := instance.GetOptions()
-	if currentOptions.Model != originalOptions.Model {
-		t.Error("Options should not change when setting nil options")
-	}
-}
-
 func TestGetProxy(t *testing.T) {
 	globalSettings := &config.InstancesConfig{
 		LogsDir: "/tmp/test",
@@ -317,58 +258,6 @@ func TestUnmarshalJSON(t *testing.T) {
 	}
 }
 
-func TestUnmarshalJSON_PartialOptions(t *testing.T) {
-	jsonData := `{
-		"name": "test-instance",
-		"running": false,
-		"options": {
-			"model": "/path/to/model.gguf"
-		}
-	}`
-
-	var inst instance.Process
-	err := json.Unmarshal([]byte(jsonData), &inst)
-	if err != nil {
-		t.Fatalf("JSON unmarshal failed: %v", err)
-	}
-
-	opts := inst.GetOptions()
-	if opts.Model != "/path/to/model.gguf" {
-		t.Errorf("Expected model '/path/to/model.gguf', got %q", opts.Model)
-	}
-
-	// Note: Defaults are NOT applied during unmarshaling
-	// They should only be applied by NewInstance or SetOptions
-	if opts.AutoRestart != nil {
-		t.Error("Expected AutoRestart to be nil (no defaults applied during unmarshal)")
-	}
-}
-
-func TestUnmarshalJSON_NoOptions(t *testing.T) {
-	jsonData := `{
-		"name": "test-instance",
-		"running": false
-	}`
-
-	var inst instance.Process
-	err := json.Unmarshal([]byte(jsonData), &inst)
-	if err != nil {
-		t.Fatalf("JSON unmarshal failed: %v", err)
-	}
-
-	if inst.Name != "test-instance" {
-		t.Errorf("Expected name 'test-instance', got %q", inst.Name)
-	}
-	if inst.Running {
-		t.Error("Expected running to be false")
-	}
-
-	opts := inst.GetOptions()
-	if opts != nil {
-		t.Error("Expected options to be nil when not provided in JSON")
-	}
-}
-
 func TestCreateInstanceOptionsValidation(t *testing.T) {
 	tests := []struct {
 		name          string
@@ -377,13 +266,6 @@ func TestCreateInstanceOptionsValidation(t *testing.T) {
 		expectedMax   int
 		expectedDelay int
 	}{
-		{
-			name:          "nil values",
-			maxRestarts:   nil,
-			restartDelay:  nil,
-			expectedMax:   0, // Should remain nil, but we can't easily test nil in this structure
-			expectedDelay: 0,
-		},
 		{
 			name:          "valid positive values",
 			maxRestarts:   testutil.IntPtr(10),
@@ -424,20 +306,16 @@ func TestCreateInstanceOptionsValidation(t *testing.T) {
 			instance := instance.NewInstance("test", globalSettings, options)
 			opts := instance.GetOptions()
 
-			if tt.maxRestarts != nil {
+			if opts.MaxRestarts == nil {
-				if opts.MaxRestarts == nil {
+				t.Error("Expected MaxRestarts to be set")
-					t.Error("Expected MaxRestarts to be set")
+			} else if *opts.MaxRestarts != tt.expectedMax {
-				} else if *opts.MaxRestarts != tt.expectedMax {
+				t.Errorf("Expected MaxRestarts %d, got %d", tt.expectedMax, *opts.MaxRestarts)
-					t.Errorf("Expected MaxRestarts %d, got %d", tt.expectedMax, *opts.MaxRestarts)
-				}
 			}
 
-			if tt.restartDelay != nil {
+			if opts.RestartDelay == nil {
-				if opts.RestartDelay == nil {
+				t.Error("Expected RestartDelay to be set")
-					t.Error("Expected RestartDelay to be set")
+			} else if *opts.RestartDelay != tt.expectedDelay {
-				} else if *opts.RestartDelay != tt.expectedDelay {
+				t.Errorf("Expected RestartDelay %d, got %d", tt.expectedDelay, *opts.RestartDelay)
-					t.Errorf("Expected RestartDelay %d, got %d", tt.expectedDelay, *opts.RestartDelay)
-				}
 			}
 		})
 	}

pkg/instance/lifecycle.go

@@ -4,6 +4,7 @@ import (
 	"context"
 	"fmt"
 	"log"
+	"net/http"
 	"os/exec"
 	"runtime"
 	"syscall"
@@ -30,6 +31,9 @@ func (i *Process) Start() error {
 		i.restarts = 0
 	}
 
+	// Initialize last request time to current time when starting
+	i.lastRequestTime.Store(i.timeProvider.Now().Unix())
+
 	// Create log files
 	if err := i.logger.Create(); err != nil {
 		return fmt.Errorf("failed to create log files: %w", err)
@@ -140,6 +144,74 @@ func (i *Process) Stop() error {
 	return nil
 }
 
+func (i *Process) WaitForHealthy(timeout int) error {
+	if !i.Running {
+		return fmt.Errorf("instance %s is not running", i.Name)
+	}
+
+	if timeout <= 0 {
+		timeout = 30 // Default to 30 seconds if no timeout is specified
+	}
+
+	ctx, cancel := context.WithTimeout(context.Background(), time.Duration(timeout)*time.Second)
+	defer cancel()
+
+	// Get instance options to build the health check URL
+	opts := i.GetOptions()
+	if opts == nil {
+		return fmt.Errorf("instance %s has no options set", i.Name)
+	}
+
+	// Build the health check URL directly
+	host := opts.Host
+	if host == "" {
+		host = "localhost"
+	}
+	healthURL := fmt.Sprintf("http://%s:%d/health", host, opts.Port)
+
+	// Create a dedicated HTTP client for health checks
+	client := &http.Client{
+		Timeout: 5 * time.Second, // 5 second timeout per request
+	}
+
+	// Helper function to check health directly
+	checkHealth := func() bool {
+		req, err := http.NewRequestWithContext(ctx, "GET", healthURL, nil)
+		if err != nil {
+			return false
+		}
+
+		resp, err := client.Do(req)
+		if err != nil {
+			return false
+		}
+		defer resp.Body.Close()
+
+		return resp.StatusCode == http.StatusOK
+	}
+
+	// Try immediate check first
+	if checkHealth() {
+		return nil // Instance is healthy
+	}
+
+	// If immediate check failed, start polling
+	ticker := time.NewTicker(1 * time.Second)
+	defer ticker.Stop()
+
+	for {
+		select {
+		case <-ctx.Done():
+			return fmt.Errorf("timeout waiting for instance %s to become healthy after %d seconds", i.Name, timeout)
+		case <-ticker.C:
+			if checkHealth() {
+				return nil // Instance is healthy
+			}
+			// Continue polling
+		}
+	}
+}
+
 func (i *Process) monitorProcess() {
 	defer func() {
 		i.mu.Lock()

pkg/instance/timeout.go

@@ -0,0 +1,28 @@
+package instance
+
+// UpdateLastRequestTime updates the last request access time for the instance via proxy
+func (i *Process) UpdateLastRequestTime() {
+	i.mu.Lock()
+	defer i.mu.Unlock()
+
+	lastRequestTime := i.timeProvider.Now().Unix()
+	i.lastRequestTime.Store(lastRequestTime)
+}
+
+func (i *Process) ShouldTimeout() bool {
+	i.mu.RLock()
+	defer i.mu.RUnlock()
+
+	if !i.Running || i.options.IdleTimeout == nil || *i.options.IdleTimeout <= 0 {
+		return false
+	}
+
+	// Check if the last request time exceeds the idle timeout
+	lastRequest := i.lastRequestTime.Load()
+	idleTimeoutMinutes := *i.options.IdleTimeout
+
+	// Convert timeout from minutes to seconds for comparison
+	idleTimeoutSeconds := int64(idleTimeoutMinutes * 60)
+
+	return (i.timeProvider.Now().Unix() - lastRequest) > idleTimeoutSeconds
+}

pkg/instance/timeout_test.go

@@ -0,0 +1,195 @@
+package instance_test
+
+import (
+	"llamactl/pkg/backends/llamacpp"
+	"llamactl/pkg/config"
+	"llamactl/pkg/instance"
+	"llamactl/pkg/testutil"
+	"sync/atomic"
+	"testing"
+	"time"
+)
+
+// MockTimeProvider implements TimeProvider for testing
+type MockTimeProvider struct {
+	currentTime atomic.Int64 // Unix timestamp
+}
+
+func NewMockTimeProvider(t time.Time) *MockTimeProvider {
+	m := &MockTimeProvider{}
+	m.currentTime.Store(t.Unix())
+	return m
+}
+
+func (m *MockTimeProvider) Now() time.Time {
+	return time.Unix(m.currentTime.Load(), 0)
+}
+
+func (m *MockTimeProvider) SetTime(t time.Time) {
+	m.currentTime.Store(t.Unix())
+}
+
+// Timeout-related tests
+
+func TestUpdateLastRequestTime(t *testing.T) {
+	globalSettings := &config.InstancesConfig{
+		LogsDir: "/tmp/test",
+	}
+
+	options := &instance.CreateInstanceOptions{
+		LlamaServerOptions: llamacpp.LlamaServerOptions{
+			Model: "/path/to/model.gguf",
+		},
+	}
+
+	inst := instance.NewInstance("test-instance", globalSettings, options)
+
+	// Test that UpdateLastRequestTime doesn't panic
+	inst.UpdateLastRequestTime()
+}
+
+func TestShouldTimeout_NotRunning(t *testing.T) {
+	globalSettings := &config.InstancesConfig{
+		LogsDir: "/tmp/test",
+	}
+
+	idleTimeout := 1 // 1 minute
+	options := &instance.CreateInstanceOptions{
+		IdleTimeout: &idleTimeout,
+		LlamaServerOptions: llamacpp.LlamaServerOptions{
+			Model: "/path/to/model.gguf",
+		},
+	}
+
+	inst := instance.NewInstance("test-instance", globalSettings, options)
+
+	// Instance is not running, should not timeout regardless of configuration
+	if inst.ShouldTimeout() {
+		t.Error("Non-running instance should never timeout")
+	}
+}
+
+func TestShouldTimeout_NoTimeoutConfigured(t *testing.T) {
+	globalSettings := &config.InstancesConfig{
+		LogsDir: "/tmp/test",
+	}
+
+	tests := []struct {
+		name        string
+		idleTimeout *int
+	}{
+		{"nil timeout", nil},
+		{"zero timeout", testutil.IntPtr(0)},
+		{"negative timeout", testutil.IntPtr(-5)},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			options := &instance.CreateInstanceOptions{
+				IdleTimeout: tt.idleTimeout,
+				LlamaServerOptions: llamacpp.LlamaServerOptions{
+					Model: "/path/to/model.gguf",
+				},
+			}
+
+			inst := instance.NewInstance("test-instance", globalSettings, options)
+			// Simulate running state
+			inst.Running = true
+
+			if inst.ShouldTimeout() {
+				t.Errorf("Instance with %s should not timeout", tt.name)
+			}
+		})
+	}
+}
+
+func TestShouldTimeout_WithinTimeLimit(t *testing.T) {
+	globalSettings := &config.InstancesConfig{
+		LogsDir: "/tmp/test",
+	}
+
+	idleTimeout := 5 // 5 minutes
+	options := &instance.CreateInstanceOptions{
+		IdleTimeout: &idleTimeout,
+		LlamaServerOptions: llamacpp.LlamaServerOptions{
+			Model: "/path/to/model.gguf",
+		},
+	}
+
+	inst := instance.NewInstance("test-instance", globalSettings, options)
+	inst.Running = true
+
+	// Update last request time to now
+	inst.UpdateLastRequestTime()
+
+	// Should not timeout immediately
+	if inst.ShouldTimeout() {
+		t.Error("Instance should not timeout when last request was recent")
+	}
+}
+
+func TestShouldTimeout_ExceedsTimeLimit(t *testing.T) {
+	globalSettings := &config.InstancesConfig{
+		LogsDir: "/tmp/test",
+	}
+
+	idleTimeout := 1 // 1 minute
+	options := &instance.CreateInstanceOptions{
+		IdleTimeout: &idleTimeout,
+		LlamaServerOptions: llamacpp.LlamaServerOptions{
+			Model: "/path/to/model.gguf",
+		},
+	}
+
+	inst := instance.NewInstance("test-instance", globalSettings, options)
+	inst.Running = true
+
+	// Use MockTimeProvider to simulate old last request time
+	mockTime := NewMockTimeProvider(time.Now())
+	inst.SetTimeProvider(mockTime)
+
+	// Set last request time to now
+	inst.UpdateLastRequestTime()
+
+	// Advance time by 2 minutes (exceeds 1 minute timeout)
+	mockTime.SetTime(time.Now().Add(2 * time.Minute))
+
+	if !inst.ShouldTimeout() {
+		t.Error("Instance should timeout when last request exceeds idle timeout")
+	}
+}
+
+func TestTimeoutConfiguration_Validation(t *testing.T) {
+	globalSettings := &config.InstancesConfig{
+		LogsDir: "/tmp/test",
+	}
+
+	tests := []struct {
+		name            string
+		inputTimeout    *int
+		expectedTimeout int
+	}{
+		{"default value when nil", nil, 0},
+		{"positive value", testutil.IntPtr(10), 10},
+		{"zero value", testutil.IntPtr(0), 0},
+		{"negative value gets corrected", testutil.IntPtr(-5), 0},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			options := &instance.CreateInstanceOptions{
+				IdleTimeout: tt.inputTimeout,
+				LlamaServerOptions: llamacpp.LlamaServerOptions{
+					Model: "/path/to/model.gguf",
+				},
+			}
+
+			inst := instance.NewInstance("test-instance", globalSettings, options)
+			opts := inst.GetOptions()
+
+			if opts.IdleTimeout == nil || *opts.IdleTimeout != tt.expectedTimeout {
+				t.Errorf("Expected IdleTimeout %d, got %v", tt.expectedTimeout, opts.IdleTimeout)
+			}
+		})
+	}
+}

pkg/manager/manager.go

@@ -10,6 +10,7 @@ import (
 	"path/filepath"
 	"strings"
 	"sync"
+	"time"
 )
 
 // InstanceManager defines the interface for managing instances of the llama server.
@@ -31,20 +32,48 @@ type instanceManager struct {
 	instances       map[string]*instance.Process
 	ports           map[int]bool
 	instancesConfig config.InstancesConfig
+
+	// Timeout checker
+	timeoutChecker *time.Ticker
+	shutdownChan   chan struct{}
+	shutdownDone   chan struct{}
+	isShutdown     bool
 }
 
 // NewInstanceManager creates a new instance of InstanceManager.
 func NewInstanceManager(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,
+
+		timeoutChecker: time.NewTicker(time.Duration(instancesConfig.TimeoutCheckInterval) * time.Minute),
+		shutdownChan:   make(chan struct{}),
+		shutdownDone:   make(chan struct{}),
 	}
 
 	// Load existing instances from disk
 	if err := im.loadInstances(); err != nil {
 		log.Printf("Error loading instances: %v", err)
 	}
+
+	// Start the timeout checker goroutine after initialization is complete
+	go func() {
+		defer close(im.shutdownDone)
+
+		for {
+			select {
+			case <-im.timeoutChecker.C:
+				im.checkAllTimeouts()
+			case <-im.shutdownChan:
+				return // Exit goroutine on shutdown
+			}
+		}
+	}()
+
 	return im
 }
 
@@ -94,6 +123,27 @@ func (im *instanceManager) Shutdown() {
 	im.mu.Lock()
 	defer im.mu.Unlock()
 
+	// Check if already shutdown
+	if im.isShutdown {
+		return
+	}
+	im.isShutdown = true
+
+	// Signal the timeout checker to stop
+	close(im.shutdownChan)
+
+	// Release lock temporarily to wait for goroutine
+	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()
+	}
+
 	var wg sync.WaitGroup
 	wg.Add(len(im.instances))
 

pkg/manager/operations.go

@@ -27,10 +27,6 @@ func (im *instanceManager) CreateInstance(name string, options *instance.CreateI
 		return nil, fmt.Errorf("instance options cannot be nil")
 	}
 
-	if len(im.instances) >= im.instancesConfig.MaxInstances && im.instancesConfig.MaxInstances != -1 {
-		return nil, fmt.Errorf("maximum number of instances (%d) reached", im.instancesConfig.MaxInstances)
-	}
-
 	name, err := validation.ValidateInstanceName(name)
 	if err != nil {
 		return nil, err
@@ -44,6 +40,11 @@ func (im *instanceManager) CreateInstance(name string, options *instance.CreateI
 	im.mu.Lock()
 	defer im.mu.Unlock()
 
+	// Check max instances limit after acquiring the lock
+	if len(im.instances) >= im.instancesConfig.MaxInstances && im.instancesConfig.MaxInstances != -1 {
+		return nil, fmt.Errorf("maximum number of instances (%d) reached", im.instancesConfig.MaxInstances)
+	}
+
 	// Check if instance with this name already exists
 	if im.instances[name] != nil {
 		return nil, fmt.Errorf("instance with name %s already exists", name)

pkg/manager/timeout.go

@@ -0,0 +1,26 @@
+package manager
+
+import "log"
+
+func (im *instanceManager) checkAllTimeouts() {
+	im.mu.RLock()
+	var timeoutInstances []string
+
+	// Identify instances that should timeout
+	for _, inst := range im.instances {
+		if inst.ShouldTimeout() {
+			timeoutInstances = append(timeoutInstances, inst.Name)
+		}
+	}
+	im.mu.RUnlock() // Release read lock before calling StopInstance
+
+	// Stop the timed-out instances
+	for _, name := range timeoutInstances {
+		log.Printf("Instance %s has timed out, stopping it", name)
+		if _, err := im.StopInstance(name); err != nil {
+			log.Printf("Error stopping instance %s: %v", name, err)
+		} else {
+			log.Printf("Instance %s stopped successfully", name)
+		}
+	}
+}

pkg/server/handlers.go

@@ -28,7 +28,23 @@ func NewHandler(im manager.InstanceManager, cfg config.AppConfig) *Handler {
 	}
 }
 
-// HelpHandler godoc
+// VersionHandler godoc
+// @Summary Get llamactl version
+// @Description Returns the version of the llamactl command
+// @Tags version
+// @Security ApiKeyAuth
+// @Produces text/plain
+// @Success 200 {string} string "Version information"
+// @Failure 500 {string} string "Internal Server Error"
+// @Router /version [get]
+func (h *Handler) VersionHandler() http.HandlerFunc {
+	return func(w http.ResponseWriter, r *http.Request) {
+		w.Header().Set("Content-Type", "text/plain")
+		fmt.Fprintf(w, "Version: %s\nCommit: %s\nBuild Time: %s\n", h.cfg.Version, h.cfg.CommitHash, h.cfg.BuildTime)
+	}
+}
+
+// LlamaServerHelpHandler godoc
 // @Summary Get help for llama server
 // @Description Returns the help text for the llama server command
 // @Tags server
@@ -37,7 +53,7 @@ func NewHandler(im manager.InstanceManager, cfg config.AppConfig) *Handler {
 // @Success 200 {string} string "Help text"
 // @Failure 500 {string} string "Internal Server Error"
 // @Router /server/help [get]
-func (h *Handler) HelpHandler() http.HandlerFunc {
+func (h *Handler) LlamaServerHelpHandler() http.HandlerFunc {
 	return func(w http.ResponseWriter, r *http.Request) {
 		helpCmd := exec.Command("llama-server", "--help")
 		output, err := helpCmd.CombinedOutput()
@@ -50,7 +66,7 @@ func (h *Handler) HelpHandler() http.HandlerFunc {
 	}
 }
 
-// VersionHandler godoc
+// LlamaServerVersionHandler godoc
 // @Summary Get version of llama server
 // @Description Returns the version of the llama server command
 // @Tags server
@@ -59,7 +75,7 @@ func (h *Handler) HelpHandler() http.HandlerFunc {
 // @Success 200 {string} string "Version information"
 // @Failure 500 {string} string "Internal Server Error"
 // @Router /server/version [get]
-func (h *Handler) VersionHandler() http.HandlerFunc {
+func (h *Handler) LlamaServerVersionHandler() http.HandlerFunc {
 	return func(w http.ResponseWriter, r *http.Request) {
 		versionCmd := exec.Command("llama-server", "--version")
 		output, err := versionCmd.CombinedOutput()
@@ -72,7 +88,7 @@ func (h *Handler) VersionHandler() http.HandlerFunc {
 	}
 }
 
-// ListDevicesHandler godoc
+// LlamaServerListDevicesHandler godoc
 // @Summary List available devices for llama server
 // @Description Returns a list of available devices for the llama server
 // @Tags server
@@ -81,7 +97,7 @@ func (h *Handler) VersionHandler() http.HandlerFunc {
 // @Success 200 {string} string "List of devices"
 // @Failure 500 {string} string "Internal Server Error"
 // @Router /server/devices [get]
-func (h *Handler) ListDevicesHandler() http.HandlerFunc {
+func (h *Handler) LlamaServerListDevicesHandler() http.HandlerFunc {
 	return func(w http.ResponseWriter, r *http.Request) {
 		listCmd := exec.Command("llama-server", "--list-devices")
 		output, err := listCmd.CombinedOutput()
@@ -100,7 +116,7 @@ func (h *Handler) ListDevicesHandler() http.HandlerFunc {
 // @Tags instances
 // @Security ApiKeyAuth
 // @Produces json
-// @Success 200 {array} Instance "List of instances"
+// @Success 200 {array} instance.Process "List of instances"
 // @Failure 500 {string} string "Internal Server Error"
 // @Router /instances [get]
 func (h *Handler) ListInstances() http.HandlerFunc {
@@ -127,8 +143,8 @@ func (h *Handler) ListInstances() http.HandlerFunc {
 // @Accept json
 // @Produces json
 // @Param name path string true "Instance Name"
-// @Param options body CreateInstanceOptions true "Instance configuration options"
+// @Param options body instance.CreateInstanceOptions true "Instance configuration options"
-// @Success 201 {object} Instance "Created instance details"
+// @Success 201 {object} instance.Process "Created instance details"
 // @Failure 400 {string} string "Invalid request body"
 // @Failure 500 {string} string "Internal Server Error"
 // @Router /instances/{name} [post]
@@ -168,7 +184,7 @@ func (h *Handler) CreateInstance() http.HandlerFunc {
 // @Security ApiKeyAuth
 // @Produces json
 // @Param name path string true "Instance Name"
-// @Success 200 {object} Instance "Instance details"
+// @Success 200 {object} instance.Process "Instance details"
 // @Failure 400 {string} string "Invalid name format"
 // @Failure 500 {string} string "Internal Server Error"
 // @Router /instances/{name} [get]
@@ -202,8 +218,8 @@ func (h *Handler) GetInstance() http.HandlerFunc {
 // @Accept json
 // @Produces json
 // @Param name path string true "Instance Name"
-// @Param options body CreateInstanceOptions true "Instance configuration options"
+// @Param options body instance.CreateInstanceOptions true "Instance configuration options"
-// @Success 200 {object} Instance "Updated instance details"
+// @Success 200 {object} instance.Process "Updated instance details"
 // @Failure 400 {string} string "Invalid name format"
 // @Failure 500 {string} string "Internal Server Error"
 // @Router /instances/{name} [put]
@@ -242,7 +258,7 @@ func (h *Handler) UpdateInstance() http.HandlerFunc {
 // @Security ApiKeyAuth
 // @Produces json
 // @Param name path string true "Instance Name"
-// @Success 200 {object} Instance "Started instance details"
+// @Success 200 {object} instance.Process "Started instance details"
 // @Failure 400 {string} string "Invalid name format"
 // @Failure 500 {string} string "Internal Server Error"
 // @Router /instances/{name}/start [post]
@@ -275,7 +291,7 @@ func (h *Handler) StartInstance() http.HandlerFunc {
 // @Security ApiKeyAuth
 // @Produces json
 // @Param name path string true "Instance Name"
-// @Success 200 {object} Instance "Stopped instance details"
+// @Success 200 {object} instance.Process "Stopped instance details"
 // @Failure 400 {string} string "Invalid name format"
 // @Failure 500 {string} string "Internal Server Error"
 // @Router /instances/{name}/stop [post]
@@ -308,7 +324,7 @@ func (h *Handler) StopInstance() http.HandlerFunc {
 // @Security ApiKeyAuth
 // @Produces json
 // @Param name path string true "Instance Name"
-// @Success 200 {object} Instance "Restarted instance details"
+// @Success 200 {object} instance.Process "Restarted instance details"
 // @Failure 400 {string} string "Invalid name format"
 // @Failure 500 {string} string "Internal Server Error"
 // @Router /instances/{name}/restart [post]
@@ -456,6 +472,9 @@ func (h *Handler) ProxyToInstance() http.HandlerFunc {
 			proxyPath = "/" + proxyPath
 		}
 
+		// 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
@@ -556,8 +575,23 @@ func (h *Handler) OpenAIProxy() http.HandlerFunc {
 		}
 
 		if !inst.Running {
-			http.Error(w, "Instance is not running", http.StatusServiceUnavailable)
+			if inst.GetOptions().OnDemandStart != nil && *inst.GetOptions().OnDemandStart {
-			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
+				}
+
+			} else {
+				http.Error(w, "Instance is not running", http.StatusServiceUnavailable)
+				return
+			}
 		}
 
 		proxy, err := inst.GetProxy()
@@ -566,6 +600,9 @@ func (h *Handler) OpenAIProxy() http.HandlerFunc {
 			return
 		}
 
+		// Update last request time for the instance
+		inst.UpdateLastRequestTime()
+
 		// Recreate the request body from the bytes we read
 		r.Body = io.NopCloser(bytes.NewReader(bodyBytes))
 		r.ContentLength = int64(len(bodyBytes))

pkg/server/routes.go

@@ -8,7 +8,7 @@ import (
 	"github.com/go-chi/cors"
 	httpSwagger "github.com/swaggo/http-swagger"
 
-	_ "llamactl/docs"
+	_ "llamactl/apidocs"
 	"llamactl/webui"
 )
 
@@ -42,10 +42,12 @@ func SetupRouter(handler *Handler) *chi.Mux {
 			r.Use(authMiddleware.AuthMiddleware(KeyTypeManagement))
 		}
 
+		r.Get("/version", handler.VersionHandler()) // Get server version
+
 		r.Route("/server", func(r chi.Router) {
-			r.Get("/help", handler.HelpHandler())
+			r.Get("/help", handler.LlamaServerHelpHandler())
-			r.Get("/version", handler.VersionHandler())
+			r.Get("/version", handler.LlamaServerVersionHandler())
-			r.Get("/devices", handler.ListDevicesHandler())
+			r.Get("/devices", handler.LlamaServerListDevicesHandler())
 		})
 
 		// Instance management endpoints

webui/src/App.tsx

@@ -7,6 +7,7 @@ import SystemInfoDialog from "./components/SystemInfoDialog";
 import { type CreateInstanceOptions, type Instance } from "@/types/instance";
 import { useInstances } from "@/contexts/InstancesContext";
 import { useAuth } from "@/contexts/AuthContext";
+import { ThemeProvider } from "@/contexts/ThemeContext";
 
 function App() {
   const { isAuthenticated, isLoading: authLoading } = useAuth();
@@ -42,44 +43,50 @@ function App() {
   // Show loading spinner while checking auth
   if (authLoading) {
     return (
-      <div className="min-h-screen bg-gray-50 flex items-center justify-center">
+      <ThemeProvider>
-        <div className="text-center">
+        <div className="min-h-screen bg-background flex items-center justify-center">
-          <div className="animate-spin rounded-full h-8 w-8 border-b-2 border-blue-600 mx-auto mb-4"></div>
+          <div className="text-center">
-          <p className="text-gray-600">Loading...</p>
+            <div className="animate-spin rounded-full h-8 w-8 border-b-2 border-primary mx-auto mb-4"></div>
+            <p className="text-muted-foreground">Loading...</p>
+          </div>
         </div>
-      </div>
+      </ThemeProvider>
     );
   }
 
   // Show login dialog if not authenticated
   if (!isAuthenticated) {
     return (
-      <div className="min-h-screen bg-gray-50">
+      <ThemeProvider>
-        <LoginDialog open={true} />
+        <div className="min-h-screen bg-background">
-      </div>
+          <LoginDialog open={true} />
+        </div>
+      </ThemeProvider>
     );
   }
 
   // Show main app if authenticated
   return (
-    <div className="min-h-screen bg-gray-50">
+    <ThemeProvider>
-      <Header onCreateInstance={handleCreateInstance} onShowSystemInfo={handleShowSystemInfo} />
+      <div className="min-h-screen bg-background">
-      <main className="container mx-auto max-w-4xl px-4 py-8">
+        <Header onCreateInstance={handleCreateInstance} onShowSystemInfo={handleShowSystemInfo} />
-        <InstanceList editInstance={handleEditInstance} />
+        <main className="container mx-auto max-w-4xl px-4 py-8">
-      </main>
+          <InstanceList editInstance={handleEditInstance} />
+        </main>
 
-      <InstanceDialog
+        <InstanceDialog
-        open={isInstanceModalOpen}
+          open={isInstanceModalOpen}
-        onOpenChange={setIsInstanceModalOpen}
+          onOpenChange={setIsInstanceModalOpen}
-        onSave={handleSaveInstance}
+          onSave={handleSaveInstance}
-        instance={editingInstance}
+          instance={editingInstance}
-      />
+        />
 
-      <SystemInfoDialog
+        <SystemInfoDialog
-        open={isSystemInfoModalOpen}
+          open={isSystemInfoModalOpen}
-        onOpenChange={setIsSystemInfoModalOpen}
+          onOpenChange={setIsSystemInfoModalOpen}
-      />
+        />
-    </div>
+      </div>
+    </ThemeProvider>
   );
 }
 

webui/src/__tests__/App.test.tsx

@@ -55,6 +55,21 @@ describe('App Component - Critical Business Logic Only', () => {
     vi.mocked(instancesApi.list).mockResolvedValue(mockInstances)
     window.sessionStorage.setItem('llamactl_management_key', 'test-api-key-123')
     global.fetch = vi.fn(() => Promise.resolve(new Response(null, { status: 200 })))
+    
+    // Mock window.matchMedia for dark mode functionality
+    Object.defineProperty(window, 'matchMedia', {
+      writable: true,
+      value: vi.fn().mockImplementation((query: string) => ({
+        matches: false,
+        media: query,
+        onchange: null,
+        addListener: vi.fn(),
+        removeListener: vi.fn(),
+        addEventListener: vi.fn(),
+        removeEventListener: vi.fn(),
+        dispatchEvent: vi.fn(),
+      })),
+    })
   })
 
   afterEach(() => {

webui/src/components/Header.tsx

@@ -1,6 +1,7 @@
 import { Button } from "@/components/ui/button";
-import { HelpCircle, LogOut } from "lucide-react";
+import { HelpCircle, LogOut, Moon, Sun } from "lucide-react";
 import { useAuth } from "@/contexts/AuthContext";
+import { useTheme } from "@/contexts/ThemeContext";
 
 interface HeaderProps {
   onCreateInstance: () => void;
@@ -9,6 +10,7 @@ interface HeaderProps {
 
 function Header({ onCreateInstance, onShowSystemInfo }: HeaderProps) {
   const { logout } = useAuth();
+  const { theme, toggleTheme } = useTheme();
 
   const handleLogout = () => {
     if (confirm("Are you sure you want to logout?")) {
@@ -17,10 +19,10 @@ function Header({ onCreateInstance, onShowSystemInfo }: HeaderProps) {
   };
 
   return (
-    <header className="bg-white border-b border-gray-200">
+    <header className="bg-card border-b border-border">
       <div className="container mx-auto max-w-4xl px-4 py-4">
         <div className="flex items-center justify-between">
-          <h1 className="text-2xl font-bold text-gray-900">
+          <h1 className="text-2xl font-bold text-foreground">
             Llamactl Dashboard
           </h1>
 
@@ -29,6 +31,16 @@ function Header({ onCreateInstance, onShowSystemInfo }: HeaderProps) {
               Create Instance
             </Button>
 
+            <Button
+              variant="outline"
+              size="icon"
+              onClick={toggleTheme}
+              data-testid="theme-toggle-button"
+              title={`Switch to ${theme === 'light' ? 'dark' : 'light'} mode`}
+            >
+              {theme === 'light' ? <Moon className="h-4 w-4" /> : <Sun className="h-4 w-4" />}
+            </Button>
+
             <Button
               variant="outline"
               size="icon"

webui/src/components/InstanceList.tsx

@@ -18,8 +18,8 @@ function InstanceList({ editInstance }: InstanceListProps) {
     return (
       <div className="flex items-center justify-center py-12" aria-label="Loading">
         <div className="text-center">
-          <div className="animate-spin rounded-full h-8 w-8 border-b-2 border-blue-600 mx-auto mb-4"></div>
+          <div className="animate-spin rounded-full h-8 w-8 border-b-2 border-primary mx-auto mb-4"></div>
-          <p className="text-gray-600">Loading instances...</p>
+          <p className="text-muted-foreground">Loading instances...</p>
         </div>
       </div>
     )
@@ -28,7 +28,7 @@ function InstanceList({ editInstance }: InstanceListProps) {
   if (error) {
     return (
       <div className="text-center py-12">
-        <div className="text-red-600 mb-4">
+        <div className="text-destructive mb-4">
           <p className="text-lg font-semibold">Error loading instances</p>
           <p className="text-sm">{error}</p>
         </div>
@@ -39,15 +39,15 @@ function InstanceList({ editInstance }: InstanceListProps) {
   if (instances.length === 0) {
     return (
       <div className="text-center py-12">
-        <p className="text-gray-600 text-lg mb-2">No instances found</p>
+        <p className="text-foreground text-lg mb-2">No instances found</p>
-        <p className="text-gray-500 text-sm">Create your first instance to get started</p>
+        <p className="text-muted-foreground text-sm">Create your first instance to get started</p>
       </div>
     )
   }
 
   return (
     <div className="space-y-4">
-      <h2 className="text-xl font-semibold text-gray-900 mb-6">
+      <h2 className="text-xl font-semibold text-foreground mb-6">
         Instances ({instances.length})
       </h2>
       

webui/src/components/LogDialog.tsx

@@ -11,6 +11,7 @@ import {
   DialogTitle,
 } from '@/components/ui/dialog'
 import { Badge } from '@/components/ui/badge'
+import { instancesApi } from '@/lib/api'
 import { 
   RefreshCw, 
   Download, 
@@ -46,48 +47,44 @@ const LogsDialog: React.FC<LogsDialogProps> = ({
   const refreshIntervalRef = useRef<NodeJS.Timeout | null>(null)
 
   // Fetch logs function
-  const fetchLogs = async (lines?: number) => {
+  const fetchLogs = React.useCallback(
-    if (!instanceName) return
+    async (lines?: number) => {
-    
+      if (!instanceName) return
-    setLoading(true)
-    setError(null)
-    
-    try {
-      const params = lines ? `?lines=${lines}` : ''
-      const response = await fetch(`/api/v1/instances/${instanceName}/logs${params}`)
       
-      if (!response.ok) {
+      setLoading(true)
-        throw new Error(`Failed to fetch logs: ${response.status}`)
+      setError(null)
+      
+      try {
+        const logText = await instancesApi.getLogs(instanceName, lines)
+        setLogs(logText)
+        
+        // Auto-scroll to bottom
+        setTimeout(() => {
+          if (logContainerRef.current) {
+            logContainerRef.current.scrollTop = logContainerRef.current.scrollHeight
+          }
+        }, 100)
+      } catch (err) {
+        setError(err instanceof Error ? err.message : 'Failed to fetch logs')
+      } finally {
+        setLoading(false)
       }
-      
+    },
-      const logText = await response.text()
+    [instanceName]
-      setLogs(logText)
+  )
-      
-      // Auto-scroll to bottom
-      setTimeout(() => {
-        if (logContainerRef.current) {
-          logContainerRef.current.scrollTop = logContainerRef.current.scrollHeight
-        }
-      }, 100)
-    } catch (err) {
-      setError(err instanceof Error ? err.message : 'Failed to fetch logs')
-    } finally {
-      setLoading(false)
-    }
-  }
 
   // Initial load when dialog opens
   useEffect(() => {
     if (open && instanceName) {
-      fetchLogs(lineCount)
+      void fetchLogs(lineCount)
     }
-  }, [open, instanceName])
+  }, [open, instanceName, fetchLogs, lineCount])
 
   // Auto-refresh effect
   useEffect(() => {
     if (autoRefresh && isRunning && open) {
       refreshIntervalRef.current = setInterval(() => {
-        fetchLogs(lineCount)
+        void fetchLogs(lineCount)
       }, 2000) // Refresh every 2 seconds
     } else {
       if (refreshIntervalRef.current) {
@@ -101,7 +98,7 @@ const LogsDialog: React.FC<LogsDialogProps> = ({
         clearInterval(refreshIntervalRef.current)
       }
     }
-  }, [autoRefresh, isRunning, open, lineCount])
+  }, [autoRefresh, isRunning, open, lineCount, fetchLogs])
 
   // Copy logs to clipboard
   const copyLogs = async () => {
@@ -135,7 +132,7 @@ const LogsDialog: React.FC<LogsDialogProps> = ({
 
   // Apply new line count
   const applyLineCount = () => {
-    fetchLogs(lineCount)
+    void fetchLogs(lineCount)
     setShowSettings(false)
   }
 
@@ -198,7 +195,7 @@ const LogsDialog: React.FC<LogsDialogProps> = ({
               <Button
                 variant="outline"
                 size="sm"
-                onClick={() => fetchLogs(lineCount)}
+                onClick={() => void fetchLogs(lineCount)}
                 disabled={loading}
               >
                 {loading ? (
@@ -290,7 +287,7 @@ const LogsDialog: React.FC<LogsDialogProps> = ({
           <div className="flex items-center gap-2 w-full">
             <Button
               variant="outline"
-              onClick={copyLogs}
+              onClick={() => void copyLogs()}
               disabled={!logs}
             >
               {copied ? (

webui/src/components/SystemInfoDialog.tsx

@@ -19,6 +19,15 @@ import {
 } from 'lucide-react'
 import { serverApi } from '@/lib/api'
 
+// Helper to get version from environment
+const getAppVersion = (): string => {
+  try {
+    return (import.meta.env as Record<string, string>).VITE_APP_VERSION || 'unknown'
+  } catch {
+    return 'unknown'
+  }
+}
+
 interface SystemInfoModalProps {
   open: boolean
   onOpenChange: (open: boolean) => void
@@ -109,9 +118,20 @@ const SystemInfoDialog: React.FC<SystemInfoModalProps> = ({
             </div>
           ) : systemInfo ? (
             <div className="space-y-6">
-              {/* Version Section */}
+              {/* Llamactl Version Section */}
               <div className="space-y-3">
-                <h3 className="font-semibold">Version</h3>
+                <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">

webui/src/components/ZodFormField.tsx

@@ -7,8 +7,8 @@ import { getFieldType, basicFieldsConfig } from '@/lib/zodFormUtils'
 
 interface ZodFormFieldProps {
   fieldKey: keyof CreateInstanceOptions
-  value: any
+  value: string | number | boolean | string[] | undefined
-  onChange: (key: keyof CreateInstanceOptions, value: any) => void
+  onChange: (key: keyof CreateInstanceOptions, value: string | number | boolean | string[] | undefined) => void
 }
 
 const ZodFormField: React.FC<ZodFormFieldProps> = ({ fieldKey, value, onChange }) => {
@@ -18,7 +18,7 @@ const ZodFormField: React.FC<ZodFormFieldProps> = ({ fieldKey, value, onChange }
   // Get type from Zod schema
   const fieldType = getFieldType(fieldKey)
 
-  const handleChange = (newValue: any) => {
+  const handleChange = (newValue: string | number | boolean | string[] | undefined) => {
     onChange(fieldKey, newValue)
   }
 
@@ -29,7 +29,7 @@ const ZodFormField: React.FC<ZodFormFieldProps> = ({ fieldKey, value, onChange }
           <div className="flex items-center space-x-2">
             <Checkbox
               id={fieldKey}
-              checked={value || false}
+              checked={typeof value === 'boolean' ? value : false}
               onCheckedChange={(checked) => handleChange(checked)}
             />
             <Label htmlFor={fieldKey} className="text-sm font-normal">
@@ -51,10 +51,14 @@ const ZodFormField: React.FC<ZodFormFieldProps> = ({ fieldKey, value, onChange }
             <Input
               id={fieldKey}
               type="number"
-              value={value || ''}
+              step="any" // This allows decimal numbers
+              value={typeof value === 'string' || typeof value === 'number' ? value : ''}
               onChange={(e) => {
                 const numValue = e.target.value ? parseFloat(e.target.value) : undefined
-                handleChange(numValue)
+                // Only update if the parsed value is valid or the input is empty
+                if (e.target.value === '' || (numValue !== undefined && !isNaN(numValue))) {
+                  handleChange(numValue)
+                }
               }}
               placeholder={config.placeholder}
             />
@@ -101,7 +105,7 @@ const ZodFormField: React.FC<ZodFormFieldProps> = ({ fieldKey, value, onChange }
             <Input
               id={fieldKey}
               type="text"
-              value={value || ''}
+              value={typeof value === 'string' || typeof value === 'number' ? value : ''}
               onChange={(e) => handleChange(e.target.value || undefined)}
               placeholder={config.placeholder}
             />

webui/src/contexts/ThemeContext.tsx

@@ -0,0 +1,54 @@
+import { createContext, useContext, useEffect, useState, type ReactNode } from "react";
+
+type Theme = "light" | "dark";
+
+interface ThemeContextType {
+  theme: Theme;
+  toggleTheme: () => void;
+}
+
+const ThemeContext = createContext<ThemeContextType | undefined>(undefined);
+
+interface ThemeProviderProps {
+  children: ReactNode;
+}
+
+export function ThemeProvider({ children }: ThemeProviderProps) {
+  const [theme, setTheme] = useState<Theme>(() => {
+    const stored = localStorage.getItem("theme");
+    if (stored === "light" || stored === "dark") {
+      return stored;
+    }
+    return window.matchMedia("(prefers-color-scheme: dark)").matches ? "dark" : "light";
+  });
+
+  useEffect(() => {
+    const root = document.documentElement;
+    
+    if (theme === "dark") {
+      root.classList.add("dark");
+    } else {
+      root.classList.remove("dark");
+    }
+    
+    localStorage.setItem("theme", theme);
+  }, [theme]);
+
+  const toggleTheme = () => {
+    setTheme(prevTheme => prevTheme === "light" ? "dark" : "light");
+  };
+
+  return (
+    <ThemeContext.Provider value={{ theme, toggleTheme }}>
+      {children}
+    </ThemeContext.Provider>
+  );
+}
+
+export function useTheme() {
+  const context = useContext(ThemeContext);
+  if (context === undefined) {
+    throw new Error("useTheme must be used within a ThemeProvider");
+  }
+  return context;
+}

webui/src/lib/zodFormUtils.ts

@@ -1,8 +1,7 @@
-import type { CreateInstanceOptions} from '@/schemas/instanceOptions';
+import { type CreateInstanceOptions, getAllFieldKeys } from '@/schemas/instanceOptions'
-import { getAllFieldKeys } from '@/schemas/instanceOptions'
 
 // Only define the basic fields we want to show by default
-export const basicFieldsConfig: Record<string, { 
+export const basicFieldsConfig: Record<string, {
   label: string
   description?: string
   placeholder?: string
@@ -22,6 +21,15 @@ export const basicFieldsConfig: Record<string, {
     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',

webui/src/schemas/instanceOptions.ts

@@ -6,6 +6,8 @@ export const CreateInstanceOptionsSchema = z.object({
   auto_restart: z.boolean().optional(),
   max_restarts: z.number().optional(),
   restart_delay: z.number().optional(),
+  idle_timeout: z.number().optional(),
+  on_demand_start: z.boolean().optional(),
 
   // Common params
   verbose_prompt: z.boolean().optional(),
@@ -14,12 +16,12 @@ export const CreateInstanceOptionsSchema = z.object({
   cpu_mask: z.string().optional(),
   cpu_range: z.string().optional(),
   cpu_strict: z.number().optional(),
-  priority: 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(),
-  priority_batch: z.number().optional(),
+  prio_batch: z.number().optional(),
   poll_batch: z.number().optional(),
   ctx_size: z.number().optional(),
   predict: z.number().optional(),
@@ -82,7 +84,7 @@ export const CreateInstanceOptionsSchema = z.object({
   seed: z.number().optional(),
   sampling_seq: z.string().optional(),
   ignore_eos: z.boolean().optional(),
-  temperature: z.number().optional(),
+  temp: z.number().optional(),
   top_k: z.number().optional(),
   top_p: z.number().optional(),
   min_p: z.number().optional(),
@@ -109,7 +111,7 @@ export const CreateInstanceOptionsSchema = z.object({
   json_schema: z.string().optional(),
   json_schema_file: z.string().optional(),
 
-  // Server/Example-specific params
+  // Example-specific params
   no_context_shift: z.boolean().optional(),
   special: z.boolean().optional(),
   no_warmup: z.boolean().optional(),
@@ -149,8 +151,6 @@ export const CreateInstanceOptionsSchema = z.object({
   no_prefill_assistant: z.boolean().optional(),
   slot_prompt_similarity: z.number().optional(),
   lora_init_without_apply: z.boolean().optional(),
-
-  // Speculative decoding params
   draft_max: z.number().optional(),
   draft_min: z.number().optional(),
   draft_p_min: z.number().optional(),

webui/src/vite-env.d.ts

@@ -0,0 +1,13 @@
+/// <reference types="vite/client" />
+
+declare global {
+  interface ImportMetaEnv {
+    readonly VITE_APP_VERSION?: string
+  }
+
+  interface ImportMeta {
+    readonly env: ImportMetaEnv
+  }
+}
+
+export {}

webui/tsconfig.json

@@ -18,8 +18,9 @@
     "baseUrl": ".",
     "paths": {
       "@/*": ["./src/*"]
-    }
+    },
+    "types": ["vite/client"]
   },
-  "include": ["src"],
+  "include": ["src", "src/vite-env.d.ts"],
   "references": [{ "path": "./tsconfig.node.json" }]
 }