MCP-Weather

A Model Context Protocol (MCP) server that connects to the Open-Meteo API through either streamable-http (default) or sse. This server provides weather data to AI clients using the MCP interface.

Configuration

The following configuration can be set with env variables:

Env var	Default value	Description
HOST	0.0.0.0	Address to listen to.
PORT	8080	Port to listen on.
TRANSPORT	streamable-http	Transport used by the mcp server. Must be streamable-httpor sse.

Requirements

It is recommended to use the provided Dockerimage, which requires only Docker to be installed.

If you want to install from source the following packages are required:

• Python 3.13
• uv

Installation

(Recommended) Using the Dockerimage

    docker run -p 8080:8080 ghcr.io/melvinkl/mcp-weather/server:latest

Using source

1. Clone the repository

    git clone https://github.com/MelvinKl/mcp-weather.git

2. Install the dependencies

    uv sync --dev

3. Runs the server

    uv run python src/main.py

Testing

To run all tests with coverage:

make test

This will:

• Run all tests using pytest
• Check for linting issues with flake8
• Format code with black
• Generate coverage reports in HTML and XML formats

Test coverage is maintained at 100% for all source code (exceeds 80% requirement).

License

Apache 2

Acknowledgments

• Built with the Model Context Protocol
• Uses data from the OpenMeteo API