Plugwise MCP is a TypeScript-based Model Context Protocol (MCP) server used to integrate the Plugwise smart home system into MCP-enabled AI clients. It provides automatic network discovery, device control, temperature management, energy monitoring, and multi-gateway management capabilities.
Purpose This server is designed to let AI clients (such as Cline, Cursor, VS Code Copilot, Windsurf, etc.) communicate with Plugwise gateways via standard input/output (stdio), thereby controlling smart home devices through natural language or programmatically.
Main capabilities
- AI agent mode: Enables natural language control through a built-in AI agent.
- JSON-RPC support: Provides a programmatic API for scripts and automation.
- Automatic network scan: Discovers all Plugwise gateways on the network.
- Credential management: Securely stores gateway passwords from a .env file.
- Device control: Controls thermostats, switches, and smart plugs.
- Temperature management: Sets temperatures, presets, and schedules.
- Energy monitoring: Reads power consumption and sensor data.
- Multi-gateway support: Manages multiple gateways simultaneously.
- Real-time updates: Retrieves current device status and measurements.
Usage Installation:
- Install globally via npm (recommended): npm install -g plugwise-mcp
- Or use npx directly (no installation required): npx plugwise-mcp
- Install from source: Requires Node.js 17 or higher, as well as a Plugwise gateway (Adam, Anna, Smile P1, or Stretch).
Starting the server:
- Standard MCP server mode (15+ dedicated tools): Run plugwise-mcp after installation; when running from source, use the corresponding command.
- AI agent mode (single natural language tool): Use the manage_plugwise tool.
Adding to an MCP client: The server can work with any MCP client that supports stdio transport. Configuration varies by client:
- Cline: Edit cline_mcp_settings.json.
- Cursor: Edit .cursor/mcp.json (project-level) or ~/.cursor/mcp.json (globaloffice).
- VS Code Copilot: edit .vscode/mcp.json (project-level) or user settings (global).
- Windsurf: edit ~/.codeium/windsurf/mcp_config.json.
Environment variables: The server reads the gateway password from environment variables, which can be provided via the env field of the MCP configuration or a .env file. Security note: when using the env field of the MCP configuration, credentials are securely passed to the server process; using a .env file is recommended (usually excluded from version control).
Main tools and features:
- Network discovery: connect to the Plugwise gateway.
- Device management: retrieve all devices and their current status.
- Climate control: set thermostat temperature setpoints, change preset modes, turn switches/sockets on/off.
- Gateway management: set_gateway_mode (home, away, vacation), set_dhw_mode (domestic hot water mode: auto, boost, comfort, off), set_regulation_mode (heating regulation mode), delete_notification (clear gateway notifications), reboot_gateway (reboot the gateway, use with caution).
- Resources: plugwise://devices provides access to the current status of all devices.
- Prompts: setup_guide provides comprehensive step-by-step setup instructions.
Testing: Provides a comprehensive read-only test suite, including server health checks, MCP protocol initialization, network scanning, gateway connection and information retrieval, device status reading, resources, and prompts. The tests only perform read operations and do not change device status.
Supported devices:
- Gateways: Adam (OpenTherm-capable smart home hub, thermostat control, underfloor heating), Anna (standalone thermostat gateway), Smile P1 (energy monitoring gateway: electricity, gas, solar), Stretch (legacy hub for connecting Circle smart sockets).
- Connected devices: Jip (motion sensor with illuminance detection), Lisa (radiator valve, hub required), Tom/Floor (underfloor heating controller), Koen (radiator valve, requires relaying via Plug), Plug (smart socket with power monitoring, Zigbee), Aqara Plug (third-party Zigbee smart socket), Circle (legacy Circle/Circle+ socket, only via Stretch).
Use cases Suitable for users who want to integrate the Plugwise smart home system into AI assistants or automation scripts, enabling features such as natural language control, energy monitoring, and multi-gateway management.
Limitations
- Requires Node.js 17 or higher.
- Requires Plugwise gateway hardware.
- Network security: Plugwise uses HTTP Basic Auth (not HTTPS); it is recommended to place the gateway on a secure local network, use a VPN for remote access, and consider using a separate VLAN for IoT devices.
- The API has full control over the heating system and access must be restricted.
- The project is based on the python-plugwise library, and its architecture pattern is inspired by sonos-ts-mcp.
- Current version 1.0.2, under the MIT license.
Documentation: Provides a setup guide, MCP server documentation, network scanning guide, testing documentation, and more.