MCP

A linear, one-concept-per-lesson path through the Model Context Protocol (MCP): building MCP servers first (tools, resources, prompts), then MCP clients, then wiring MCP tools into a real LangChain/Gemini agent. 26 lessons · 3 tiers.

Prerequisites: Completion of LangChain & Agent Building, or already comfortable with @tool, bind_tools, and the manual tool-calling loop. MCP tools are a server-hosted version of the same idea, a name, a description, and an argument schema, just served over a protocol instead of imported as a Python function.

Lessons use Google's Gemini free tier (gemini-3.5-flash-lite), the same model as the other three courses, once an LLM enters the picture starting at Lesson 15. Several lessons also lean on the MCP Inspector, a Node-based dev tool run with npx @modelcontextprotocol/inspector, to browse a server's tools, resources, and prompts without writing a client first.

Built on MCP, the protocol this course teaches.