---
id: ha-mcp-dev
version: "8.2.0.dev2334"
license: MIT
license_treatment: permissive
maintenance: active
---

# ha-mcp-dev — Home Assistant MCP Server - Complete control of Home Assistant through MCP

License: permissive · Maintenance: active · Downloads: 146.3K/mo

## What it is and what it does

ha-mcp-dev is a Model Context Protocol server that bridges AI assistants (Claude, ChatGPT, etc.) to Home Assistant smart home systems. It translates natural language requests into Home Assistant API calls, allowing an AI to query device states, trigger automations, and control smart home devices without the user writing code. The package is designed to run either as a Home Assistant custom component (in-process, recommended), as a standalone HTTP server via Docker or PyPI, or over stdio for local connections.

The server handles authentication through Home Assistant tokens, webhooks, or optional OIDC providers, and exposes Home Assistant's full service and state APIs to the AI model. It depends on fastmcp for the MCP protocol implementation, httpx for HTTP communication, pydantic for data validation, cryptography for secure connections, and several utility libraries (python-dotenv, truststore, tzdata, packaging). The package is actively maintained, supports Python 3.13 and 3.14, and carries no known security vulnerabilities.

Use it for:

- Let an AI assistant control lights, thermostats, and locks in your home using conversational commands.
- Query Home Assistant device states and sensor readings through an MCP-connected AI model.
- Trigger automations and scenes from an AI chat interface without manual Home Assistant UI interaction.
- Integrate Home Assistant into Claude Desktop or ChatGPT as a smart home control tool.
- Run a secure remote MCP server for AI assistants on different networks via webhook or reverse proxy.

## Worth the install?

AI-flagged interpretation of the facts above — verify before relying.

An MCP server that lets AI assistants control Home Assistant smart homes through natural language, querying device states, executing services, and managing automations.

Yes, if you use Home Assistant and want AI assistants to control your smart home. The custom component install is straightforward, maintenance is active, and there are no known vulnerabilities. The dev version tag suggests this is a pre-release; verify stability for your use case before production deployment. MIT license is permissive.

## Install

pip install ha-mcp-dev
uv add ha-mcp-dev
poetry add ha-mcp-dev

## Installing ha-mcp-dev

Before you install: Low friction install as a pure Python wheel. Active maintenance with a recent release and 4381 GitHub stars. Requires Python 3.13 or 3.14 only; nine runtime dependencies are all standard ecosystem packages (fastmcp, httpx, pydantic, cryptography, etc.).

License in practice: MIT license (permissive) — you can use, modify, and distribute this package freely with minimal restrictions.

Quickstart:

pip install ha-mcp-dev

import ha_mcp
# Configure via environment variables (HOMEASSISTANT_URL, HOMEASSISTANT_TOKEN)
# or use the Home Assistant custom component for in-process setup

Requires Python 3.13 or 3.14; requires a running Home Assistant instance and a long-lived access token or webhook URL to connect.

Verify before relying:

- Whether the dev version (8.2.0.dev2334) is stable enough for production use or intended for testing only.
- Performance characteristics when controlling many devices or handling concurrent AI requests.
- Whether all nine runtime dependencies are required or if some are optional for specific features.

## Package facts

- License: MIT (permissive)
- Python support: supports_current
- Install friction: low
- Maintenance: active
- Downloads: 146.3K/month (top 15,000 on PyPI)
- Known vulnerabilities: none known

## Tags

home assistant mcp server, ai smart home control, natural language home automation, home assistant integration ai, mcp protocol home assistant, smart home voice commands, home assistant api wrapper, home-automation, mcp-protocol, ai-integration

[View on SkillFed](https://skillfed.io/packages/ha-mcp-dev) · [View on PyPI](https://pypi.org/project/ha-mcp-dev/)