Chatbots answer. Agents act. An AI agent can read a document, search your files, run a calculation, organize a folder, and chain those steps together to finish real work. This book teaches you to build agents yourself, in Python, from a blank file to working tools, with every project explained line by line in beginner-friendly language.
Jake Morrison starts with what an agent really is, then sets up your Python workspace and the essential language skills agent builders actually use. You connect to a model API, build your first agent loop, and learn the two ideas everything else stands on: tools the model can call, and structured outputs you can rely on.
Inside, you will learn how to:
• Build an agent loop that reasons, acts, and checks its own work
• Connect Python to a model API safely and cheaply
• Define tools and structured outputs your code can trust
• Give an agent conversation state and memory
• Build a calculator agent and a document assistant
• Plan multi-step work with LangGraph
• Control safety, permissions, and API costs from day one
The project sequence keeps raising the bar: a calculator agent, a document assistant, a research agent, a file organizer, a structured data processor, and a final multi-tool workflow assistant that combines everything. Each project is complete, runnable, and small enough to understand fully, so you finish with skills instead of copy-pasted mysteries.
This book is for Python beginners, no-code tinkerers ready to level up, students, and professionals who want agents they control rather than subscriptions they rent. If you can write a simple script or are willing to learn in the first three chapters, you are ready.
Every project guards against the classic beginner traps: runaway loops, surprise API bills, and agents with more permissions than judgment. Build your first agent this week, and keep building.
