Build AI Apps with Python: The ReAct Agent Pattern — Think Act Observe | Episode 18

A chatbot responds once. An agent loops until done. The ReAct pattern: Think, Act, Observe — the foundation of every AI agent. We build a ReAct agent from scratch. Two tools — weather lookup and calculator. Claude reasons about the question, picks a tool, executes it, observes the result, and loops until it has the answer. Single-step and multi-step reasoning, all visible in the trace. Student code: https://github.com/GoCelesteAI/build-ai-apps-python/tree/main/episode18 What You'll Learn: • The ReAct pattern: Think, Act, Observe loop • Agent vs chatbot — why the while loop matters • Claude tool use with tool definitions and schemas • Tool dispatcher pattern — mapping names to functions • stop_reason to detect when the agent is done • Feeding tool results back as messages • Single-step tool use (weather lookup) • Multi-step reasoning (compare two cities)