Allows controlling Home Assistant smart home devices via MCP, including calling services (e.g., switch lights, adjust climate), getting entity states, and looking up service parameters through the Home Assistant REST API.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@Home Assistant MCP Server turn on the living room light"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
English | 中文
通过 MCP (Model Context Protocol) 让 AI Agent 直接控制你的 Home Assistant 智能家居设备。
零外部依赖 — 只需要 Node.js 18+。
Node.js 18+
正在运行的 Home Assistant 实例
Home Assistant 长期访问令牌（Long-Lived Access Token）
Related MCP server: Home Assistant MCP
工具 | 功能 | 必需参数 | 可选参数 |
| 调用 HA 服务（开关灯、调温度等） | domain, service | url, token, body |
| 获取实体状态 | — | url, token, entity_id |
| 查询服务参数信息 | — | domain, service |
所有工具都支持通过参数 url 和 token 传入 HA 凭证，也支持通过环境变量 HA_URL 和 HA_TOKEN 设置默认值。
git clone https://github.com/weyand138-netizen/Home-assistant-AI-agent-mcp.git
cd Home-assistant-AI-agent-mcp
# 直接运行（无需 npm install）
node ha-mcp-server.mjs编辑 claude_desktop_config.json：
{
"mcpServers": {
"ha": {
"command": "node",
"args": ["路径/ha-mcp-server.mjs"],
"env": {
"HA_URL": "http://192.168.1.100:8123",
"HA_TOKEN": "你的长期访问令牌"
}
}
}
}在任何支持 MCP stdio 传输的客户端中，按相同方式配置即可。也可以不设置环境变量，在对话中直接传参：
"把客厅灯打开，HA 地址是 http://192.168.1.100:8123，令牌是 xxxxx"
# 启动服务器
node ha-mcp-server.mjs
# 也可以用 MCP Inspector 调试
npx @modelcontextprotocol/inspector node ha-mcp-server.mjsAI Agent → MCP stdio (JSON-RPC 2.0) → ha-mcp-server.mjs
→ Node.js fetch（无 CORS 限制）
→ HA REST API（port 8123）
→ 智能家居设备内置 32 个 domain、100+ 个服务的参数信息。包含 homeassistant、light、switch、climate、shopping_list、todo、scene、automation 等常用领域。
如需更新字典，可在 HA 中执行 curl http://your-ha:8123/api/services，按格式整理后替换 ha-service-dict.mjs。
MIT
我只是初学者，这个做得可能不太好，欢迎指正 🙏
This server cannot be deployed
Automate 1,000+ services from any MCP-compatible AI agent: build Applets, run actions and queries.
Discover and call AI agents via MCP. Supports A2A agents and platform agents with async tasks.
AI agent registry — search, discover, register, and connect agents via MCP.
The OpenRouter for tools. One MCP connection gives any AI agent 254 hosted tools, pay per call.
