Home Assistant MCP

[](https://smithery.ai/server/@guilhermelirio/homeassistant-mpc) # Home Assistant MCP This is an MCP (Model Context Protocol) server that provides tools for integration with Home Assistant, allowing AI assistants to interact with smart home devices. ## Setup ### Installing via Smithery To install Home Assistant Integration for Claude Desktop automatically via [Smithery](https://smithery.ai/server/@guilhermelirio/homeassistant-mpc): ```bash npx -y @smithery/cli install @guilhermelirio/homeassistant-mpc --client claude ``` ### Manual Installation 1. Install dependencies: ```bash npm install ``` 2. Configure environment variables: - Obtain a Long-Lived Access Token from your Home Assistant - Configure the environment variables in a `.env` file at the project root: ``` HOME_ASSISTANT_URL=http://your-home-assistant:8123 HOME_ASSISTANT_TOKEN=your_token_here ``` ## Starting the server ```bash npm start ``` For development: ```bash npm run dev ``` ## Available Tools ### Home Assistant API Verification - **homeassistant_api**: Verifies if the Home Assistant API is online - Parameters: none - Example: `homeassistant_api()` ### Get Entity State - **homeassistant_get_state**: Gets the current state of a Home Assistant entity - Parameters: `entity_id` (entity ID) - Example: `homeassistant_get_state("light.living_room")` ### Call Services - **homeassistant_call_service**: Calls a service for a Home Assistant entity - Parameters: - `entity_id` (entity ID) - `domain` (service domain) - `service` (service to call) - Example: `homeassistant_call_service("light.living_room", "light", "turn_on")` ## Development This project uses TypeScript and the MCP SDK library to create tools that can be accessed by language models such as Claude, GPT, etc. To add new tools: 1. Create a module in `src/tools/[category]` 2. Implement your tools following the existing pattern 3. Register the module in `src/index.ts` ## Contributions Contributions are welcome! Feel free to open issues or submit pull requests. ## License This project is licensed under the MIT License - see the LICENSE file for details.