Weather MCP Server

A Model Context Protocol (MCP) server that provides weather information for different regions around the world.

Overview

This MCP server uses the wttr.in API to fetch weather data and serves it through a standardized MCP interface. It allows AI models and other clients to request weather information for any region by name.

Features

• Fetch weather information for any city or region worldwide
• Uses the free wttr.in API (no API key required)
• Implemented as a FastMCP server for easy integration with AI models
• Docker support for easy deployment
• Minimal dependencies for fast startup and low resource usage

Requirements

• Python 3.10+
• Only 2 dependencies:
  • requests - HTTP client for wttr.in API
  • fastmcp - MCP framework (includes uvicorn and starlette)

Installation

1. Clone the repository:

   git clone https://github.com/hydavinci/weather_mcp.git
   cd weather_mcp
   

2. Install dependencies using uv (recommended):

   uv sync
   

   Or using pip:

   pip install requests fastmcp
   

Usage

Running the server locally

HTTP mode (recommended for web clients):

# PowerShell
$env:TRANSPORT="http"; python src/server.py

# Bash/Linux
TRANSPORT=http python src/server.py

stdio mode (for local MCP clients):

python src/server.py

The server will start on port 8081 by default for HTTP mode.

Using Docker

Build the Docker image:

docker build -t weather-mcp-server .

Run the container:

docker run -p 8081:8081 weather-mcp-server

API Usage

The server exposes a single MCP tool:

get_weather

Fetches weather information for the specified region.

Parameters:

• region_name (string, required): The name of the city or region to get weather for.

Returns:

• JSON object with comprehensive weather information from wttr.in

Example weather data includes:

• Current conditions and temperature
• 3-day forecast with hourly details
• Astronomical data (sunrise, sunset, moon phase)
• Weather descriptions and codes
• Humidity, pressure, wind information
• UV index and visibility

Testing

You can test the weather functionality directly by running:

python src/weather_fetcher.py

This will fetch and display weather information for Suzhou (default city).

Environment Variables

• TRANSPORT: Set to "http" for HTTP mode, otherwise defaults to stdio mode
• PORT: Port for HTTP mode (default: 8081)
• SERVER_TOKEN: Optional server token for stdio mode

Project Structure

weather_mcp/
├── src/
│   ├── server.py           # Main MCP server
│   ├── weather_fetcher.py  # Weather API client
│   └── middleware.py       # HTTP middleware
├── pyproject.toml          # Project configuration
├── Dockerfile             # Container configuration
├── smithery.yaml          # Smithery deployment config
└── README.md              # This file

License

See the LICENSE file for details.

Build the Docker image:

docker build -t weather-mcp-server .

Run the container:

docker run -p 8000:8000 weather-mcp-server

API Usage

The server exposes a single MCP handler:

get_weather

Fetches weather information for the specified region.

Parameters:

• region_name (string, required): The name of the city or region to get weather for.

Returns:

• JSON object with weather information from wttr.in

Example:

# Example client code
from mcp.client import MCPClient

async with MCPClient("http://localhost:8000") as client:
    weather_data = await client.get_weather(region_name="London")
    print(weather_data)

Testing

You can test the weather functionality directly by running:

python weather_fetcher.py

This will fetch and display weather information for Suzhou (default city).

License

See the LICENSE file for details.