WeatherMCP

This project is an MCP server built as a hands-on exercise. It provides weather data and alerts from the US National Weather Service (NWS) via a simple API using the FastMCP framework.

Features

• Weather Alerts: Get active weather alerts for any US state.
• Forecast: Retrieve weather forecasts for a specific latitude and longitude.

Requirements

• Python 3.8+
• uv (for dependency management and running)

Installation

1. Install dependencies using uv:

   uv pip install -r requirements.txt

   Or, if you use pyproject.toml:

   uv pip install -r pyproject.toml

2. (Optional) Create a virtual environment:

   uv venv .venv
   source .venv/bin/activate

Running the MCP Server

To start the server using uv:

uv pip install -e .  # if you want to install as editable, or just ensure dependencies are installed
uv python weather.py

The server will start and listen for MCP requests via stdio.

Available Tools

1. get_alerts(state: str)

• Description: Get active weather alerts for a US state.
• Argument: state — Two-letter US state code (e.g., CA, NY).
• Returns: Formatted string of current alerts or a message if none are found.

2. get_forecast(latitude: float, longitude: float)

• Description: Get weather forecast for a specific location.
• Arguments:
  • latitude — Latitude of the location (float)
  • longitude — Longitude of the location (float)
• Returns: Formatted string with the next 5 forecast periods.

Notes

• This project is for educational and hands-on purposes.
• Data is sourced from the National Weather Service API.

---

Feel free to extend or modify the project for your own experiments!