Weather MCP Server

A Model Context Protocol (MCP) server that provides weather information through Claude for Desktop. This server integrates with the National Weather Service API to provide weather forecasts and alerts for US locations. Created as part of the MIT MCP Hackathon (April 2025).

Features

get_forecast: Get detailed weather forecasts for any US location using latitude and longitude
get_alerts: Retrieve active weather alerts for any US state using state code

Requirements

Python 3.10 or higher
uv package manager
Claude for Desktop
MCP SDK 1.2.0 or higher

Installation

Ensure you have uv installed:

curl -LsSf https://astral.sh/uv/install.sh | sh

Set up the Python environment:

# Create and activate virtual environment
uv venv
source .venv/bin/activate

# Install dependencies
uv add "mcp[cli]" httpx

Configuration

Create or edit your Claude for Desktop configuration at ~/Library/Application Support/Claude/claude_desktop_config.json:

{
  "mcpServers": {
    "weather": {
      "command": "/path/to/uv",
      "args": [
        "--directory",
        "/absolute/path/to/weather",
        "run",
        "weather.py"
      ]
    }
  }
}

Replace /path/to/uv with your uv executable path (find it using which uv)
Replace /absolute/path/to/weather with the absolute path to this project directory

Usage

Start Claude for Desktop
Look for the hammer 🔨 icon to confirm the server is connected
Example queries:
- "What's the weather in Sacramento?"
- "What are the active weather alerts in Texas?"

Troubleshooting

Check Claude's logs at ~/Library/Logs/Claude/mcp*.log
Ensure all paths in the configuration are absolute
Verify the server runs independently using uv run weather.py
Check port 8001 is not in use by another process

API Reference

The server uses the National Weather Service API:

Base URL: https://api.weather.gov
Endpoints used:
- /alerts/active/area/{state} - Get alerts by state
- /points/{lat},{lon} - Get forecast grid data
- /forecast - Get detailed forecast

About the MIT MCP Hackathon

This project was created as part of a tutorial during the MIT MCP Hackathon (April 2025), which aimed to introduce developers to the Model Context Protocol and help projects/startups launch their own MCP servers. The hackathon was part of the NANDA (Networked Agents and Decentralized AI) hub initiative at MIT.

License

This project follows the original tutorial from Model Context Protocol

Screenshots