Inputs: prompts, tools, and data
One kind of argument. A prompt, a tool, a number, your own object — each becomes a variable in the agent's REPL and is described in its prompt. Nothing else to learn first.
Open in ColabJAZ is a Python framework for building LLM agents that write and run their own code. Each tutorial below is a Google Colab notebook — Google’s free hosted Python environment — so you can run a real JAZ agent in your browser and watch it reason in code.
Start at the top and work down. Each notebook builds on the one before, and every one ships with its outputs — so you can read the whole sequence before running anything.
What you’ll need:
If some of the terms on the cards are new (invoke, ReturnType, hooks, budgets), the
Concepts page defines them in plain language.
One kind of argument. A prompt, a tool, a number, your own object — each becomes a variable in the agent's REPL and is described in its prompt. Nothing else to learn first.
Open in ColabAn agent that calls agents. Why the inputs you pass stop at the call that received them, and how scope hands a tool to the whole tree.
Open in ColabTyped returns, validation, budgets and limits turn out to be one mechanism: a hook observes events and returns effects. Write your own.
Open in ColabThe LLM, the REPL and the protocol underneath it all — including what the agent is allowed to import, and how to change any of it for a single call.
Open in ColabThe homepage's program with every feature on: the penguins DataFrame and a live web search in, a typed report out, validated, logged and on a budget.
Open in ColabOnce you’re ready to build locally, follow the Installation guide.