AI agents

Two projects that turn a language model from a talker into a doer: a Python agent that picks its tools in a safe loop, then an MCP server that makes your tools available to any AI app.

OverviewTwo projects that turn a language model from a talker into a doer: a Python agent that picks its tools in a safe loop, then an MCP server that makes your tools available to any AI app.
Building an AI agent in PythonAn agent is a language model in a loop: it picks a tool to call, your program runs it, and the result goes back until it is done. We build the loop from scratch with a fake model, then see the same loop with the real Claude API.
Building an MCP server in PythonThe Model Context Protocol (MCP) is an open standard that makes your tools available to any AI app that supports it. We understand the messages (JSON-RPC), build a server with tools and resources, and test it.