What is MCP, and how do I use this?

MCP (Model Context Protocol) is an open standard that lets an AI assistant fetch live data or take actions during a conversation, instead of relying only on what it learned during training.

Each server on the front page is one such data source, exposed as a set of "tools" (for example get today's weather, or look up a company). You never call these tools yourself — you connect an MCP client to a server's web address (its endpoint), and the assistant decides on its own when a tool is needed to answer your question.

How to use these

1. Pick a server on the front page and copy its endpoint, e.g. https://allemannsdata.com/weather/mcp.
2. In your MCP client's settings, add it as a remote server (often called "Add server," "Connectors," or "MCP servers" — check your client's own documentation for the exact wording).
3. Ask the assistant something that needs that data. It calls the right tool automatically — no commands to type.

Quick-start example

Most MCP clients accept a small JSON config with this shape (this example connects the weather server):

{
  "mcpServers": {
    "weather": {
      "type": "http",
      "url": "https://allemannsdata.com/weather/mcp"
    }
  }
}

Already know the ropes?

See the full documentation for the complete endpoint list, Claude Code CLI setup, and raw .mcp.json config.