This project bridges the physical world and AI large language models by translating the Model Context Protocol (MCP) to MQTT. It enables users to control hardware, including smart home devices and robots, using natural language commands processed by LLMs, facilitating real-time AI-driven adjustments to physical parameters.

How It Works

The system leverages MQTT for robust, publish/subscribe-based communication between devices and the central controller. It fully supports the Model Context Protocol (MCP), enabling resource management and tool invocation. Commands are sent via MQTT topics, and responses are received on designated topics, allowing for flexible integration with various AI clients like Claude Desktop and Continue.

Quick Start & Requirements

Installation is supported via Python scripts for Windows, macOS, and Ubuntu/Raspberry Pi. The project relies on the uv package manager for dependency handling. A functional MQTT broker (e.g., Mosquitto) is required.

Highlighted Details

• Full support for Model Context Protocol (MCP), including resource management and tool invocation.
• Flexible configuration for multiple MQTT brokers and command definitions.
• Supports various AI clients: Claude Desktop, Continue, and Cline.
• Includes a device simulator (responder.py) for testing.

Maintenance & Community

The project is hosted on GitHub. Links to documentation, API docs, and configuration guides are provided within the README.

Licensing & Compatibility

The repository does not explicitly state a license. Compatibility for commercial use or closed-source linking is not specified.

Limitations & Caveats

The project is primarily described in Chinese, with English documentation potentially less comprehensive. No specific hardware requirements beyond MQTT connectivity are detailed, and the absence of a clear license may pose adoption challenges for commercial applications.