windy-mcp-server

Unofficial MCP server for the Windy Point Forecast API. This project is not affiliated with or endorsed by Windy.com.

An MCP (Model Context Protocol) server that exposes weather, ocean wave, and air quality forecasts to AI assistants via the Windy Point Forecast API.

Tools

get_weather_forecast

Fetch atmospheric weather data for any coordinates.

• Parameters: temperature, dewpoint, relative humidity, pressure, geopotential height, precipitation (total/snow/convective), wind, wind gusts, cloud cover (low/mid/high), cloud base, visibility, CAPE, precipitation type, weather warnings
• Models: GFS, ICON, ICON-EU, ICON-D2, AROME (France, Antilles, Réunion), NAM (CONUS/Hawaii/Alaska), HRRR (CONUS/Alaska), HRDPS
• Pressure levels: surface through 150 hPa (14 levels)
• Temperature units: Celsius or Kelvin

get_wave_forecast

Fetch ocean wave data for any coastal or open-water coordinates.

• Parameters: significant wave height, wind waves, wave power, swell 1 & 2
• Models: GFS Wave, ICON Wave, ICON-EU Wave, RDWPS (Canada)

get_air_quality_forecast

Fetch air quality and pollen forecasts for any location.

• Parameters: AQI, SO₂, dust, CO, O₃, NO₂, PM10, PM2.5, pollen (alder, birch, grass, mugwort, olive, ragweed)
• Models: CAMS (global), CAMS-EU (higher-resolution European data)

Setup

Prerequisites: Python 3.12+, uv, a Windy API key

1. Clone the repository and install dependencies:

   git clone https://github.com/yourusername/windy-mcp-server.git
   cd windy-mcp-server
   uv sync
   

2. Create a .env file with your API key:

   cp .env.example .env
   # Edit .env and set WINDY_API_KEY=your_api_key_here
   

3. Add to your MCP client config (e.g. claude_desktop_config.json):

   {
     "mcpServers": {
       "windy": {
         "command": "uv",
         "args": ["run", "python", "server/server.py"],
         "cwd": "/path/to/windy-mcp-server",
         "env": {
           "WINDY_API_KEY": "your_api_key_here"
         }
       }
     }
   }
   

License

MIT