aareguru

The Aareguru MCP Server provides tools and prompts to fetch and interpret Swiss Aare river data, including water temperature, flow, safety assessments, forecasts, and historical data. It exposes 6 MCP tools, 3 data resources accessible via aareguru:// URIs, and several prompts for daily reports, spot comparisons, and trend analyses. You can connect the MCP server to Claude or FastMCP Cloud to query the data in natural language, with Swiss German language touches in temperature descriptions and safety context. The recommended path is to use the FastMCP Cloud deployment for zero-config hosting, or run the server locally via uv for development and testing, then point your client (Claude or other MCP clients) to the server’s MCP endpoint.

Prerequisites:

• Python 3.10+ (as indicated by the project)
• uv (https://astral.sh/uv/) for local development and running MCPs
• Optional: Docker if you prefer containerized runs

Install and run locally (USB/UV workflow):

1. Install uv following the instructions in the README’s Local Installation section:

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

2. Clone the MCP repository and navigate into it:

git clone https://github.com/schlpbch/aareguru-mcp.git
cd aareguru-mcp

3. Synchronize with uv (will install dependencies if needed):

uv sync

4. Run tests (optional during development):

uv run pytest

Claude Desktop (Local) configuration example:

{
  "mcpServers": {
    "aareguru": {
      "command": "uv",
      "args": ["--directory", "/path/to/aareguru-mcp", "run", "aareguru-mcp"]
    }
  }
}

Docker deployment notes (optional):

• Follow the repository’s Docker instructions, copy .env.example to .env, and use docker-compose up -d to start the service, then verify health at http://localhost:8000/health.

If you prefer a containerized or cloud deployment, consider FastMCP Cloud as described in the Hosting section of the README.

Tips and notes:

• This MCP server is now maintained primarily in the swiss-mcp mono-repo. For current development, refer to packages/aareguru-mcp/.
• The server exposes a health endpoint for health checks when hosted (e.g., FastMCP Cloud). Use the MCP URL provided by your hosting environment.
• If you encounter environment-specific path issues, ensure the --directory path in the uv command points to the local aareguru-mcp project root.
• The server supports Prometheus metrics via /metrics when hosted, enabling monitoring of tool usage and durations.
• No authentication is required for the basic deployment; adjust security settings in your hosting platform if needed.