LiveKit Agent integrated with Home Assistant MCP Server

by longcw

22 stars
339 downloads
Not rated
GitHub

About

LiveKit Agent integrated with MCP server of Home Assistant

Details

Author
longcw
GitHub stars
22
Downloads
339
Categories
AI

- Voice control of Home Assistant devices through LiveKit Agents
- Real-time audio/video communication with AI assistants
- Access to Home Assistant entities and services via MCP
- Uses OpenAI for natural language processing

Enable the MCP Server integration in Home Assistant, set up a LiveKit agent using the LiveKit Agents framework, and configure it with your Home Assistant URL, access token, LiveKit credentials, and OpenAI API key. Then run the agent in a Python environment.

# LiveKit Agent integrated with Home Assistant MCP Server This project integrates [LiveKit Agents](https://github.com/livekit/agents) with Home Assistant's [Model Context Protocol (MCP) Server](https://www.home-assistant.io/integrations/mcp_server/), allowing voice control of your smart home through LiveKit agents. ## Overview This integration enables: - Voice control of Home Assistant devices through LiveKit Agents - Real-time audio/video communication with AI assistants - Access to Home Assistant entities and services via MCP ## Prerequisites - Home Assistant with MCP Server integration enabled - LiveKit and OpenAI settings - Python environment for running the agent ## Setup 1. Configure Home Assistant MCP Server following the [official documentation](https://www.home-assistant.io/integrations/mcp_server/) 2. Set up your LiveKit agent using the [LiveKit Agents framework](https://github.com/livekit/agents) 3. Configure the agent with your Home Assistant URL and access token ## Configuration ```bash # Example configuration LIVEKIT_URL = "wss://your-livekit-server.com" LIVEKIT_API_KEY = "your-api-key" LIVEKIT_API_SECRET = "your-api-secret" OPENAI_API_KEY HOME_ASSISTANT_URL = "http://your-homeassistant:8123/mcp_server/sse" HOME_ASSISTANT_TOKEN = "your-long-lived-access-token" ```
No reviews yet — be the first

Sign in to leave a review

Use Google, GitHub, or an email account so ratings stay tied to real people.

Email sign in

No reviews posted yet.