OpenHue MCP Server

OpenHue MCP Server

An MCP server that enables control of Philips Hue lights through Claude and other LLM interfaces using the OpenHue CLI.

Prerequisites

Bridge Setup

Before using the server, you need to set up the OpenHue CLI with your Hue Bridge:

  1. Run the setup command:
# On Linux/macOS: docker run -v "${HOME}/.openhue:/.openhue" --rm --name=openhue -it openhue/cli setup # On Windows (PowerShell): docker run -v "${env:USERPROFILE}\.openhue:/.openhue" --rm --name=openhue -it openhue/cli setup
  1. Follow the on-screen instructions:
    • The CLI will search for your Hue Bridge
    • Press the link button on your Hue Bridge when prompted
    • Wait for confirmation that the setup is complete
  2. Verify the setup by listing your lights:
# On Linux/macOS: docker run -v "${HOME}/.openhue:/.openhue" --rm --name=openhue -it openhue/cli get lights # On Windows (PowerShell): docker run -v "${env:USERPROFILE}\.openhue:/.openhue" --rm --name=openhue -it openhue/cli get lights

If you see your lights listed, the setup is complete and you're ready to use the MCP server.

Installation

  1. Clone the repository:
git clone <your-repo-url> cd claude-mcp-openhue
  1. Install dependencies:
npm install
  1. Build the project:
npm run build
  1. Run the server:
npm start

Features

This server exposes the following capabilities through MCP:

Lights Control

  • List all lights or get specific light details
  • Turn lights on/off
  • Adjust brightness
  • Set colors
  • Control color temperature

Room Control

  • List all rooms or get room details
  • Control all lights in a room together
  • Set room-wide brightness and colors

Scene Management

  • List available scenes
  • Activate scenes with different modes
  • Filter scenes by room

Usage with Claude Desktop

  1. Open your Claude Desktop configuration file:
    • macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
    • Windows: %APPDATA%\Claude\claude_desktop_config.json
  2. Add the server configuration:
{ "mcpServers": { "hue": { "command": "node", "args": ["/absolute/path/to/build/index.js"] } } }
  1. Restart Claude Desktop
  2. Look for the hammer icon to verify the server is connected

Example Commands

Once connected, you can ask Claude natural language questions like:

  • "What lights do I have in the living room?"
  • "Turn on all the lights in the kitchen"
  • "Set the bedroom lights to 50% brightness"
  • "Change the office lights to blue"
  • "Activate the 'Relaxing' scene"
  • "What scenes are available in the den?"

Available Tools

get-lights

Lists all lights or gets details for specific lights

{ lightId?: string; // Optional light ID or name room?: string; // Optional room name filter }

control-light

Controls individual lights

{ target: string; // Light ID or name action: "on" | "off"; brightness?: number; // 0-100 color?: string; // Color name temperature?: number; // 153-500 Mirek }

get-rooms

Lists all rooms or gets specific room details

{ roomId?: string; // Optional room ID or name }

control-room

Controls all lights in a room

{ target: string; // Room ID or name action: "on" | "off"; brightness?: number; color?: string; temperature?: number; }

get-scenes

Lists available scenes

{ room?: string; // Optional room name filter }

activate-scene

Activates a specific scene

{ name: string; // Scene name or ID room?: string; // Optional room name mode?: "active" | "dynamic" | "static"; }

Development

Project Structure

. ├── src/ │ └── index.ts # Main server implementation ├── build/ # Compiled JavaScript ├── package.json ├── tsconfig.json └── README.md

Building

npm run build

Running

npm start

Troubleshooting

Server Not Connecting

  1. Check that Docker is running
  2. Verify OpenHue configuration exists
  3. Check Claude Desktop logs
  4. Try running OpenHue CLI directly

Command Failures

  1. Check OpenHue CLI permissions
  2. Verify light/room/scene names
  3. Check Docker container logs
  4. Verify Hue Bridge connectivity

License

MIT License

Contributing

  1. Fork the repository
  2. Create your feature branch
  3. Commit your changes
  4. Push to the branch
  5. Create a new Pull Request