Before any step in this book makes sense, you need two things: a place to write, and something to write with. You will not type code by hand, so that something is an AI agent. This chapter sets both up on the computer you already own, and leaves you a project folder with your first file in it.
1.1.1Your own computer is enough
Your computer already runs Windows, macOS, or Linux. That is its operating system, the ground floor everything else installs onto, and any of the three builds everything in this book.
Any laptop from the last few years is enough: 8 GB of memory works, 16 GB is comfortable. Buy nothing and switch nothing, you start from what you have.
1.1.2The agent matters more than the editor
The one thing you cannot skip is an AI coding agent: the program that actually writes and changes your code. Since you never type code by hand, the agent is the engine this whole book runs on.
Two things decide how good it is: the agent itself, and the model powering it, the AI brain it thinks with. Keep an eye on both. A sharp agent on a weak model, or the reverse, will let you down.
1.1.3Install one agent and sign in
Where you and the agent work together is your choice, and it matters far less than the agent itself. Pick one of these two now:
- Want the simplest path? Install an agent-first tool like Claude Code, Codex, Cursor, or OpenCode, where the agent is the whole interface.
- More technical, or want full control? Install VS Code, a free editor from Microsoft, and run your agent inside it.
Installing is a download and a normal install, the same as any other app. Then two steps people are never warned about: you create an account, and you pay for a plan. You will run this agent constantly, so a plan that keeps up matters more than the exact tool you land on.
Hint: In 2026, Claude Code is the strongest agent for this, and it is the one this book uses. New options appear all the time, so ask your agent or search for the current best, then pick the one you feel comfortable in.
1.1.4Worth adding on day one
If you picked Claude Code, a few plugins are worth installing before you write your first prompt:
- Superpowers, a library of skills that keeps the agent from skipping steps like planning and testing.
- Ponytail, which pushes back on over-engineering and keeps the agent's solutions as small as the problem actually is.
- chrome-devtools-mcp, which lets the agent open a real browser and see your app the way a user would, instead of guessing from the code.
- This book's own skill, so your agent knows what this handbook covers and can point you at the right chapter instead of you hunting for it.
Ask your agent to find and install each one by name. As you hit a specific need later, chapter by chapter, ask it to search for a plugin for that need too, the right set grows with your project instead of being decided all at once here.
1.1.5The agent is your only real cost until you deploy
Budget for this before anything else. A usable plan starts around 20 US dollars a month, and the heavier tiers that survive a full day of building run 100 to 200.
Two doors reach the same models. One is a flat monthly plan; the other is pay-per-use billing, charged by how much text the agent reads and writes. Take the flat plan, because an agent reads whole files all day, and per-use billing is how people get a shocking bill in week one.
Free tiers exist, and they throttle you exactly when you are mid-build. Fine for an evening of trying a tool out, not something to build on.
1.1.6Meet the terminal, where commands run
A command is one line of text that tells your computer to do something, and it goes in the terminal: a plain text panel where you type that line and your computer runs it. The terminal already ships with your operating system, and VS Code has one built in, so there is nothing to install.
You will barely use it by hand. Your agent runs commands there itself and shows you the output, usually asking permission first. That prompt is not a warning, it is the agent doing what it should: telling you before it touches your machine.
1.1.7Make a folder and your first file
Create one new folder on your computer for this project; it will hold everything you build. Inside it, put a single README.md saying, in your own words, what you are building. That is your whole workspace today.
That file is not paperwork. It is the first thing your agent reads to know what this project is, and it keeps growing as the project does. Every other folder gets created by the chapter that first needs it.
1.1.8Have your agent finish the setup
Every chapter from here ends with a prompt: a block of text you copy and paste to your agent as its instructions. This is your first one. With the agent installed, open it and paste this. It checks your machine, builds your folders, and confirms everything works before you go further:
Do this now: install your agent, sign in, pay for a plan, then paste the prompt and let it set up your project folder and write your first README.
