HA MCP Server
An MCP server that exposes Home Assistant's REST and WebSocket APIs as 182 tools, enabling clients like Claude Code to control lights, climate, media, automations, and more. It also provides a status dashboard and supports running as a Home Assistant app or container.
README
HA MCP Server
An MCP server that exposes Home Assistant's REST and WebSocket APIs as 182 tools, so an MCP client — Claude Code among others — can drive an instance directly.
Lights, climate, covers, media, alarm and cameras; automations, scripts, scenes, helpers and dashboards; the area, floor, label, device and entity registries; add-ons, HACS, backups and system health; history, logbook, statistics and energy.
A status dashboard shows what the server sees and which tools have been called.
It can also delete things — automations, dashboards, helpers, users. Every tool was exercised against a live instance before release, but treat it as what it is: a broad, powerful interface to your home.

The screenshot is generated by
scripts/screenshots.pyagainst a mock of the Home Assistant API, never against a real instance — the dashboard puts the location name in the page title.
Two ways to run it
As a Home Assistant app
The easiest route if you already run Home Assistant: the app reaches it through the Supervisor, so no long-lived token has to be created, and the dashboard is available in the sidebar through ingress.
Repository: driin0/home-assistant-apps. That repository holds only the packaging — the code is here, and it ships the image published from this repository.
As a container
For a Home Assistant instance you do not control, or to run the server elsewhere:
git clone https://github.com/driin0/ha-mcp-server
cd ha-mcp-server
cp .env.sample .env # fill in HA_URL, HA_TOKEN and MCP_SECRET
docker compose up -d
The image (ghcr.io/driin0/ha-mcp-server) is published multi-architecture
(amd64, arm64), so nothing is built locally. To work on the code, replace image: with build: . in
compose.yaml.
| Variable | Default | Meaning |
|---|---|---|
HA_URL |
— | Home Assistant base URL, e.g. http://homeassistant.local:8123 |
HA_TOKEN |
— | a long-lived access token |
MCP_SECRET |
— | bearer token clients must present |
MCP_ALLOW_NO_AUTH |
false |
start without a secret — only on a trusted network |
MCP_PORT / UI_PORT |
47821 / 47822 |
ports on the host |
HA_DEFAULT_LANGUAGE |
empty | language for conversation tools |
HA_ALEXA_KEYWORDS |
echo,alexa |
media players treated as Alexa devices |
HA_REMOTE_PREFIXES |
empty | entity prefixes routed as remotes |
Connecting a client
{
"mcpServers": {
"homeassistant": {
"type": "http",
"url": "http://<host>:47821/mcp",
"headers": { "Authorization": "Bearer <MCP_SECRET>" }
}
}
}
Development
pip install -r requirements.txt
HA_URL=… HA_TOKEN=… MCP_ALLOW_NO_AUTH=true python3 server.py
The tools live in tools/, one module per area, and register themselves through
tools/_base.py. web.py serves the dashboard; stats.py keeps the call
counters behind it.
A note on the base image
The image is built on python:3.13-alpine on purpose, and its Alpine and
Python versions must stay aligned with the Home Assistant base image
(ghcr.io/home-assistant/base-python:3.13-alpineX.Y). The Home Assistant app
copies the compiled packages from this image rather than reinstalling them, and
native extensions — pydantic_core, cryptography, rpds, websockets —
only load if musl and the Python minor version match. The Dockerfile imports
them at build time so a future mismatch fails there, not at first start.
Licence
Copyright (C) 2026 Riccardo Riina (driin0)
Distributed under the GNU Affero General Public License v3.0. See LICENSE.
Trademarks
Home Assistant and its logo are trademarks of the Open Home Foundation. The Model Context Protocol and its mark belong to Anthropic. This project is independent and is not affiliated with, nor endorsed by, either.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.