Claude: Zero to Hero
Start here · module 1 of 4

How to use this track

You are here · All paths · Free plan · 10 min · Assumes nothing. This is the first file to read.

What you’ll learn: how to work through this material so it actually sticks.

If you only read one thing

This track is 60-odd short guides, grouped into stages. You don’t do all of it — you pick one of the four paths in the README and follow that. Every guide has the same shape, and every guide ends with exercises.

The exercises are the point. Reading about prompting improves your prompting roughly as much as reading about swimming improves your swimming. If you’re going to skip something, skip the deep explanations and keep the exercises — not the other way round.

The shape of every module

Every file follows the same five sections. When you’re skimming, you can jump straight to Try it.

Section What it is
What you’ll learn One sentence. If you already know it, skip the module.
Concepts The guide. Read once, properly.
Try it Exercises. Do them in a real Claude window, not in your head.
Checkpoint A self-test. If you can’t do it, reread.
Going deeper Official doc links for the parts this guide compresses.

What you need before you start

Minimum: a free Claude account at claude.ai. Stages 00 and most of 01 work on free.

For stage 02 (power user): a paid plan — Pro, Max, Team, or Enterprise. Projects, custom Skills, Cowork, connectors and file creation are paid features.

For stage 03 (Claude Code): a paid Claude plan or a Claude Console account with credits, plus a terminal.

For stages 04–06 (API): a Claude Console account with prepaid credits at platform.claude.com. Budget $5–20 to complete every exercise comfortably.

See 03-plans-and-pricing.md for the details.

How to practise so it sticks

Use a scratch project

Before module 01, create a folder on your computer called claude-practice. Everything you build in the exercises goes there. By stage 03 you’ll be pointing Claude Code at it.

Keep a prompt journal

One markdown file. Every time a prompt works well, paste it in with a one-line note on why. This is the single highest-leverage habit in the whole track — you’re building a personal prompt library, and by stage 02 you’ll convert the best entries into Skills.

Do the exercises badly, on purpose

Several exercises ask you to write a deliberately vague prompt first, then a specific one. Do the vague version. Seeing the bad output is the lesson; skipping to the good prompt teaches you nothing.

Don’t chase completeness

You will not memorise every flag and parameter. Nobody does. The goal is knowing that a capability exists and roughly where to look. The cheat sheets in 99-reference/ exist so you can forget the details safely.

A note on how fast this moves

Anthropic ships changes to Claude Code weekly and new models every few months. Specific things that will drift:

Things that will not drift much, and which is why most of this track is durable:

Weight your attention accordingly.

Try it

  1. Create your claude-practice folder.
  2. Create claude-practice/prompt-journal.md with a single heading: # Prompts that worked.
  3. Open claude.ai and confirm you can send a message.

Checkpoint

You can answer: which of the four learning paths in the README am I on, and which folders am I skipping?

Going deeper