Overview
What is mcp-weather?
mcp-weather is an MCP server example for AI agents to retrieve US weather forecasts and alerts. It integrates with the National Weather Service API and uses SSE transport for remote, decoupled access.
How to use mcp-weather?
Clone the repository, install dependencies with uv sync, then start the server with uv run server or via Docker. Connect using an MCP client (e.g., MCP Inspector) via SSE URL http://<host>:<port>/sse. The server exposes two tools: get_alerts and get_forecast. Host and port can be configured with command-line arguments.
Key features of mcp-weather
- Provides US weather forecasts and alerts
- SSE-based for remote, decoupled access
- Two tools: get_alerts and get_forecast
- Supports uv and Docker deployment
- Configurable host and port via command line
Use cases of mcp-weather
- AI agent fetching weather alerts for a US state
- AI agent retrieving forecast for specific coordinates
- Demonstrating SSE-based MCP server-client interaction
FAQ from mcp-weather
What transport does mcp-weather use?
SSE (Server-Sent Events), allowing the server to run as a remote process separate from the client.
What tools does mcp-weather provide?
Two tools: get_alerts for alerts by US state, and get_forecast for forecast by latitude/longitude.
How do I run mcp-weather?
Run with uv run server after installing dependencies, or build and run as a Docker container.
What are the system requirements?
Requires Python and uv runtime, or Docker. The examples use Ubuntu 24.04.
Is mcp-weather limited to US weather?
Yes, the tools are for US states and coordinates, using the National Weather Service API.
