A Model Context Protocol (MCP) server that enables AI assistants to interact with and control a local LedFX instance.
This MCP server provides a bridge between AI assistants (like Claude) and LedFX, allowing you to control your LED lighting setup through natural language. The server exposes LedFX's functionality as MCP tools that AI assistants can use to:
- Query device and virtual information
- List and manage LED devices and virtuals
- Apply effects to virtuals (virtual LED strips)
- Manage and activate scenes
- Create custom palettes and playlists
- Get AI-powered effect recommendations
- Get system information
- 🎨 Control LED effects through natural language
- 🔧 Manage multiple LED devices and virtuals
- 🎭 Activate pre-configured scenes
- 📊 Query device and system information
- 🔒 Type-safe TypeScript implementation
- 🏗️ Built following software design best practices
- 🎨 LedFX colors & gradients via
/api/colors(builtin + user-defined) - 🗂️ Palette management stored as user gradients in LedFX
- 🤖 AI-powered scene creation from natural language descriptions
- 📝 Playlist support for scene sequences
- 💡 Effect recommendations based on mood and description
- 📚 LedFX feature explanations - learn about any LedFX concept
- 🔄 Correct API implementation - uses virtuals (not devices) for effects
This project follows principles from:
- Separation of Concerns: Actions (I/O operations) are clearly separated from calculations (pure functions)
- Stratified Design: Clear abstraction layers (tools → client → API)
- Immutability: Data transformations use pure functions where possible
- Deep Modules: Complex LedFX API interactions hidden behind simple interfaces
- Information Hiding: Implementation details abstracted from callers
- Minimize Complexity: Each module has a single, focused responsibility
Current Status: Implemented with Comprehensive Test Suite
This MCP server is implemented against current LedFX APIs and validated with automated tests plus continuous integration checks. The implementation fixes critical API issues (virtuals vs devices) and adds advanced features like palette management, natural language scene creation, and AI-powered recommendations.
- API Corrections Applied: Effects are correctly applied to virtuals (not devices), matching current LedFX API behavior
- Comprehensive Testing: 34 tests covering unit and E2E scenarios - all passing
- CI/CD Pipeline: GitHub Actions workflow with lint, build, test, and coverage jobs
- Production Ready: Fully documented with installation guide, usage examples, and architecture docs
- API Specification - Complete LedFX API reference
- Test Specification - Comprehensive test plans
- Implementation Notes - Design decisions and technical notes
- References - LedFX resources and links
- Installation Guide - Step-by-step setup instructions
- Usage Guide - How to use all features
Status: Ready for production use with LedFX 2.1.4+
- Node.js 24 or 25
- A running LedFX instance (default:
localhost:8888) - An MCP-compatible AI assistant (e.g., Claude Desktop)
Using Docker (Recommended):
# Using docker-compose (included in this repository)
docker-compose up -d
# Or using docker run
docker run -d --name ledfx -p 8888:8888 ledfxorg/ledfx:latestUsing pip:
pip install ledfx
ledfx --host 0.0.0.0 --port 8888See the official LedFX links in REFERENCES.md for current installation docs.
# Clone the repository
git clone https://github.com/abossard/ledfx-mcp.git
cd ledfx-mcp
# Install dependencies
npm install
# Build and run in one command
npm run dev
The server connects to LedFX using the following environment variables:
LEDFX_HOST: LedFX server host (default:localhost)LEDFX_PORT: LedFX server port (default:8888)
You can set these in your MCP client configuration or as environment variables.
Add the following to your Claude Desktop configuration file:
- Claude Desktop launches this MCP server over stdio.
- Use Node.js 24 or 25.
- The path must be absolute when using
nodewithdist/index.js.
MacOS: ~/Library/Application Support/Claude/claude_desktop_config.json
Windows: %APPDATA%/Claude/claude_desktop_config.json
{
"mcpServers": {
"ledfx": {
"command": "node",
"args": ["/absolute/path/to/ledfx-mcp/dist/index.js"],
"env": {
"LEDFX_HOST": "localhost",
"LEDFX_PORT": "8888"
}
}
}
}Auto-recompile on every start (recommended for development):
Using npm run start ensures the TypeScript source is recompiled (prestart hook) each time the MCP client launches the server, so you never run stale code:
{
"mcpServers": {
"ledfx": {
"command": "npm",
"args": ["run", "start", "--prefix", "/absolute/path/to/ledfx-mcp"],
"env": {
"LEDFX_HOST": "localhost",
"LEDFX_PORT": "8888"
}
}
}
}If you prefer to run from Git without a local build, use npx as the command:
{
"mcpServers": {
"ledfx": {
"command": "npx",
"args": ["github:abossard/ledfx-mcp"],
"env": {
"LEDFX_HOST": "localhost",
"LEDFX_PORT": "8888"
}
}
}
}Once configured, you can interact with your LedFX setup through natural language:
Basic Operations:
- "List all my LED virtuals"
- "Activate the rainbow effect on my desk light virtual"
- "Show me all available scenes"
- "Clear effects from all virtuals"
Natural Language Scene Creation:
- "Create a calm ocean scene with slow blue waves"
- "Make an energetic party scene with fast rainbow colors"
- "Create a romantic scene with dim pink and purple gradients"
- "Build a focus scene with steady white light at medium brightness"
Color and Palette Management:
- "List all LedFX colors and gradients"
- "Get color or gradient 'sunset'"
- "Create a user color 'my-magenta' = #FF00FF"
- "Create a new palette called 'Sunset Vibes' with #FFA500, #FF69B4, #800080"
- "Save a playlist of my party scenes"
Effect Recommendations:
- "Recommend effects for a relaxing evening"
- "What effects work well with music?"
- "Suggest something energetic for a party"
Learning LedFX:
- "Explain what virtuals are in LedFX"
- "What's the difference between devices and virtuals?"
- "How do audio-reactive effects work?"
- "Tell me about WLED devices"
- "List all available effect types"
The server exposes 85+ MCP tools organized into categories:
| Tool | Description |
|---|---|
ledfx_get_info |
Get LedFX server information (version, features) |
ledfx_list_devices |
List all physical LED devices |
ledfx_get_device |
Get details about a specific device |
ledfx_list_virtuals |
List all virtual LED strips (includes global paused state) |
ledfx_get_virtual |
Get details about a specific virtual |
ledfx_activate_virtual |
Activate/deactivate a virtual |
ledfx_update_virtual_config |
Update virtual config (transitions, brightness, frequency range, matrix) |
ledfx_find_devices |
Trigger network device discovery |
ledfx_get_paused_state |
Check if all virtuals are globally paused |
ledfx_toggle_pause_all |
Toggle global pause on all virtuals |
ledfx_get_global_brightness |
Get global brightness value (0-1) |
ledfx_set_global_brightness |
Set global brightness for all virtuals |
ledfx_set_startup_scene |
Set scene to activate on LedFX startup |
ledfx_send_notification |
Send notification to LedFX frontend UI |
| Tool | Description |
|---|---|
ledfx_set_effect |
Apply an effect to a virtual (not device) |
ledfx_update_effect |
Update effect configuration |
ledfx_clear_effect |
Remove effects from a virtual |
ledfx_get_effect_schemas |
Get schemas for all effect types |
| Tool | Description |
|---|---|
ledfx_list_scenes |
List all available scenes |
ledfx_activate_scene |
Activate a pre-configured scene |
ledfx_create_scene |
Create new scene from current config |
ledfx_delete_scene |
Delete a saved scene |
ledfx_create_scene_from_description |
AI-powered scene creation from natural language |
| Tool | Description |
|---|---|
ledfx_list_palettes |
List all palettes stored as user gradients |
ledfx_create_palette |
Create/update a palette (stored as user gradient) |
ledfx_get_palette |
Get palette by name |
ledfx_delete_palette |
Delete a palette by name |
| Tool | Description |
|---|---|
ledfx_list_playlists |
List all playlists |
ledfx_create_playlist |
Create scene sequence playlist |
ledfx_get_playlist |
Get specific playlist |
ledfx_delete_playlist |
Delete a playlist |
| Tool | Description |
|---|---|
ledfx_list_colors |
List all colors and gradients from LedFX |
ledfx_get_color_or_gradient |
Get a specific color or gradient by ID |
ledfx_upsert_color_or_gradient |
Create/update a user color or gradient |
ledfx_delete_color_or_gradient |
Delete a user color or gradient |
ledfx_delete_user_gradients |
Delete all user-defined gradients (includes palettes) |
| Tool | Description |
|---|---|
ledfx_recommend_effects |
Get effect recommendations based on mood/description |
ledfx_explain_feature |
Get detailed explanation of any LedFX feature |
ledfx_list_features |
List all explainable features |
ledfx_list_effect_types |
List all effect types with descriptions |
| Tool | Description |
|---|---|
ledfx_get_presets |
Get presets for a virtual's effect |
ledfx_apply_preset |
Apply a preset to a virtual |
ledfx_save_preset |
Save current effect as a user preset |
ledfx_delete_preset |
Delete a preset for a virtual |
| Tool | Description |
|---|---|
ledfx_list_audio_devices |
List audio input devices |
ledfx_set_audio_device |
Set active audio device |
# Install dependencies
npm install
# Build the project
npm run build
# Build and run in one command
npm run dev
# Watch mode for development
npm run watch
# Build and run in one command
npm run dev
# Lint the code
npm run lint
# Fix linting issues
npm run lint:fixledfx-mcp/
├── src/
│ ├── index.ts # Main server entry point
│ ├── ledfx-client.ts # LedFX API client
│ └── tools.ts # MCP tool definitions and handlers
├── docs/ # Comprehensive documentation
│ ├── API_SPECIFICATION.md # LedFX API reference
│ ├── TEST_SPECIFICATION.md # Test plans and requirements
│ ├── IMPLEMENTATION_NOTES.md # Known issues and fixes needed
│ └── REFERENCES.md # LedFX resources and links
├── dist/ # Compiled JavaScript (generated)
├── docker-compose.yml # Docker setup for testing
├── package.json # Project configuration
├── tsconfig.json # TypeScript configuration
└── README.md # This file
Comprehensive documentation is available in the docs/ directory:
Complete reference for LedFX REST API endpoints (verified against LedFX 2.1.4 and upstream source):
- All endpoint paths, methods, and parameters
- Data model and orchestration caveats (virtuals, presets, scenes, playlists, blender)
- API behavior quirks and compatibility notes
Comprehensive test requirements and test cases:
- Unit test cases for all components
- Integration test scenarios
- End-to-end workflows
- Performance test criteria
- Compatibility test matrix
- Mock strategies and test fixtures
- Docker-based test environment setup
Critical analysis of current implementation vs actual API:
- Known issues and bugs (devices vs virtuals confusion)
- Required fixes before production use
- Missing features and functionality gaps
- Recommended implementation phases
- Migration path for future versions
Concise source index for API verification:
- Official LedFX documentation links
- Release/version links
- Upstream source files used to validate API behavior
The server is organized into three main layers:
- MCP Server Layer (
index.ts): Handles MCP protocol communication - Tools Layer (
tools.ts): Defines available tools and routes requests - Client Layer (
ledfx-client.ts): Abstracts LedFX HTTP API interactions
This layered architecture ensures:
- Clear separation of concerns
- Easy testing and maintenance
- Extensibility for new features
Contributions are welcome! Please ensure your code:
- Follows the existing code style
- Includes appropriate comments
- Passes linting (
npm run lint) - Builds successfully (
npm run build)
MIT License - see LICENSE file for details
- Ensure LedFX is running and accessible at the configured host/port
- Check firewall settings
- Verify the
LEDFX_HOSTandLEDFX_PORTenvironment variables
- Restart Claude Desktop after configuration changes
- Verify the path to
dist/index.jsis absolute - Check Claude Desktop logs for errors
For issues and questions: