What you are going to build

By the end of this course you will have built a support agent that reads a message, decides what to do, looks things up with tools, remembers the conversation, and stops to ask you before doing anything it cannot undo.

You will build it one piece at a time. A lesson adds a few lines, runs them, and explains what changed, so you are never handed a screen of code to decode. Every output you see on this site was captured by really running the code above it.

The whole course runs for free. No API key, no signup, no bill. Lessons 1 to 13 are plain Python. In lesson 14 you write a stand-in model, about twenty lines, that behaves like a real one everywhere it matters. Lesson 30 shows the two lines that swap in a real model when you want one.

The word agent just means a program that decides its own next step instead of following a fixed list. That is the thing you are working towards.

What you need

• Python 3.10 or later.
• Comfort with functions and dictionaries. Nothing more.

pip install langgraph langchain

Two packages. LangGraph runs your graph, and LangChain brings the models, the message types and the tools. LangGraph installs part of LangChain on its own, but not the part lesson 29 needs, so it is easier to take both now.

The one idea behind all of it

A LangGraph program is a set of Python functions that pass one dictionary between them. That is genuinely all it is.

You write the functions. You say which one runs after which. LangGraph runs them and carries the dictionary along.