Journey from a simple chat interface to a full ReAct reasoning loop. Each lab builds on the last — by Lab 7, you'll watch an AI agent think, search, calculate, and reason in real time.
Experience raw LLM I/O — text in, text out, nothing else.
One model handles summarization, code, translation, and explanation.
Discover what LLMs can't do — time, real-time data, precise math.
Watch a calculator tool eliminate LLM math errors instantly.
The LLM decides: calculator, Wikipedia, or web search?
Tools in a fixed sequence: search → summarize → report.
Thought → Action → Observation, live and streaming. The crown jewel.
Give your agent a backbone with state machines and graph visualization.