ESPHome MCP Server
Enables AI assistants to manage ESPHome devices by listing, inspecting, editing, validating, compiling, and flashing firmware over-the-air via the ESPHome WebSocket API.
README
<h1> ESPHome MCP Server <img src="images/logo.png" width="80" align="right" /> </h1>
A custom Model Context Protocol (MCP) server for managing ESPHome devices via the ESPHome Device Builder WebSocket & REST API.
It allows AI assistants (like Claude, Gemini, or custom agents) to inspect, edit, validate, compile, and flash ESPHome firmware Over-The-Air (OTA) directly from conversation context.
Features
esphome_list_devices: Lists all registered ESPHome devices, IP addresses, online statuses, board architectures, and YAML filenames.esphome_get_config: Retrieves full YAML configuration for any device (e.g.esp32-shelf.yaml).esphome_update_config: Saves updated YAML configurations directly to the ESPHome instance.esphome_validate: Validates YAML syntax & component configurations before compilation.esphome_compile: Triggers C++ compilation of device firmware.esphome_flash_ota: Flashes compiled firmware Over-The-Air (OTA) over Wi-Fi.
How It Works
The server connects directly over local WebSockets to the standard ESPHome Device Builder dashboard (port 6052) running on your network.
- No Home Assistant Add-ons Required: It bypasses third-party container layers and talks directly to ESPHome.
- No SSH / Samba Mounts Needed: Configs are read, updated, validated, and flashed over the native ESPHome WebSocket API.
Quick Start (Automated Setup)
Run the automated installer script to create the virtual environment, install dependencies, verify your ESPHome connection, and automatically register the server with your MCP client:
git clone https://github.com/saihgupr/esphome-mcp.git
cd esphome-mcp
python3 install.py
The script will prompt for your ESPHome IP/Host & Port, test the connection, create the .venv, and automatically add esphome to your mcp_config.json (Antigravity IDE, Claude Desktop, Cursor, etc.).
Note: After running
install.py, fully restart your IDE or AI Client application (Quit & Relaunch) to load the new MCP server.
Manual Configuration (Alternative)
If you prefer to configure manually, create .venv and update your MCP client configuration file (e.g. ~/.gemini/config/mcp_config.json or claude_desktop_config.json) using absolute paths:
{
"mcpServers": {
"esphome": {
"command": "/Users/your-username/path/to/esphome-mcp/.venv/bin/python",
"args": [
"/Users/your-username/path/to/esphome-mcp/server.py"
],
"env": {
"ESPHOME_HOST": "192.168.1.4",
"ESPHOME_PORT": "6052"
},
"disabled": false
}
}
}
Environment Variables
| Environment Variable | Default Value | Description |
|---|---|---|
ESPHOME_HOST |
192.168.1.4 |
IP address or hostname of your ESPHome Device Builder / Home Assistant instance. |
ESPHOME_PORT |
6052 |
Port for the ESPHome web dashboard / WebSocket API. |
Usage & Examples
Once configured in your MCP client, you can manage your ESPHome devices using natural conversation.
Example Prompts:
-
Check Device Status:
"List all my ESPHome devices and show which ones are currently online."
-
Inspect Configuration:
"Show me the configuration file for
esp32-shelf.yaml." -
Edit, Validate & OTA Update:
"Update
esp32-shelf.yamlso pressing the door button triggers a 180-degree LED hue shift for half a second. Validate the config, and if valid, flash the update OTA over Wi-Fi." -
Compile Test:
"Check
nodemcu-2.yamlfor syntax errors and compile the binary to verify the build."
Contributing
Contributions are welcome! Please submit all Pull Requests to the develop branch.
- Fork the repository
- Create a feature branch (
git checkout -b feature/awesome-feature) - Commit your changes and push to your fork
- Open a Pull Request to develop
Support & Feedback
If you encounter any issues, bugs, or have feature requests, please open an issue on GitHub.
ESPHome MCP Server is open-source and free. If you find it useful, consider giving it a star ⭐ or making a donation to support development!
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。