Welcome to the course

This course is a hands-on introduction to the Model Context Protocol. Over four modules, you will build a CLI chatbot that implements both an MCP client and an MCP server — enough working code to see every moving part of an MCP integration.

What you will build

A Python CLI chatbot backed by a document management MCP server. The server exposes tools for reading and editing in-memory documents, resources for listing and fetching content, and a slash-command prompt for reformatting documents. The client wires it all into a Claude-powered chat loop.

Most real projects implement either a client or a server. You will build both — once — so the full round-trip is visible before you specialise.

Prerequisites

• Python: comfort with functions, type hints, and async/await.
• Claude API: an Anthropic API key and prior experience calling the API.
• UV: installed via the Astral UV install guide. If you prefer not to use UV, the project also runs under plain python main.py.
• MCP background: skim the Model Context Protocol introduction if you have not already.

Course roadmap

• Module 1 — Introduction. What MCP is, what problem it solves, and how the client-server message flow works.
• Module 2 — Hands-on with MCP servers. Project setup, defining tools with the @mcp.tool decorator, and testing with the MCP Inspector.
• Module 3 — Connecting with MCP clients. Building the MCP client class, then adding resources and prompts on both sides.
• Module 4 — Assessment and wrap up. A knowledge check, then a synthesis of the three MCP primitives and when to use each.