Python MCP Server Boilerplate Documentation

The one idea behind the whole boilerplate, a map of everything inside, and what you can build with it.

Accept the GitHub invite, get the repo on your computer (with or without the command line), and know how updates reach you.

One script brings up the full stack with Docker: API, MCP server, worker, database, and admin. No Python install needed.

Point your coding agent at the repo once, so every request after this one starts with it already knowing where endpoints go.

Describe a QR code generator to your AI assistant, check the four things that matter, reload, and use it from the live docs page.

Connect the boilerplate's MCP server to Claude Code with an API key and watch Claude call the endpoint you just built.

One decorator argument makes the endpoint paid: anonymous calls get refused, users get charged per call, and the ledger records everything.

Turn the template into your product: name, brand color, plans, your endpoints, your repo. The safe order to do it in.

The exact VPS deploy we ran: one server, docker compose, HTTPS via Caddy in two lines, and your MCP server live for Claude.ai.

A map of the 30 reference docs in the repo: when you need X, read Y. Plus how to get help and what updates look like.