The official [Yeelight MCP Server](https://github.com/Yeelight/yeelight-iot-mcp) enables users to control and query their [Yeelight](https://en.yeelight.com/) smart devices using natural language, offering a seamless and efficient human-AI interaction experience.
Documentation
Yeelight MCP Server# Project Introduction
Yeelight MCP Server is an intelligent lighting control service developed based on the MCP (Model Context Protocol). This service is designed to provide a unified interface for AI assistants, intelligent agents (such as Claude Desktop, Cursor, Cherry Studio, etc.), or other clients supporting the MCP protocol, enabling efficient interaction with Yeelight Pro series smart devices.
Features
💡 Smart Device Control
🔍 Status Query Capability
🌈 Scene Mode Execution
🛠️ Easy to Extend and Integrate
🤖 AI-Friendly Design
📦 High Compatibility
Quick Start
Yeelight MCP Server supports two access methods: Official Remote MCP Server Access and Local Source Code Deployment.
📌 Prerequisites
Access Token Acquisition
HouseId Acquisition
ClientId Acquisition
🛰️ Method 1: Integrate with Yeelight Official Remote MCP Server
Official MCP Server Service Addresshttps://api.yeelight.com/apis/mcp_server/v1/mcp
🖥️ Method 2: Local Source Code Deployment
Install uv
Clone the source code
Create and activate a virtual environment
Install dependencies
Start the service
🧩 Client Configuration
Yeelight MCP Server uses the Streamable HTTP protocol for access. Below are configuration examples for some clients.