Claude Code Tutorial

20 modulesmacOS and Windows

Why this course

Claude Code is an agent: an AI that does not just make suggestions, it acts — it reads your project’s files, runs commands in the terminal, edits the code, checks the result and corrects itself until the job is done. It changes the way software is built, and the old habits are no longer enough.

This course takes you from the first installation to the workflows that are actually used in production: managing the AI’s memory, the CLAUDE.md instructions file, the extensions (skills, MCP, hooks, subagents — each explained in its own module, and all of them in the Glossary), headless automation and reusable project templates. Every module ends with links to the official documentation: they are the map for going deeper and the place to read about what’s new.

Foundations

Setup on macOS and Windows, Claude in IDEs, anatomy of the configuration files, agentic coding, context management, CLAUDE.md and Plan Mode. Modules 0–6.

Customization and extensions

Slash commands, skills, MCP servers, hooks and subagents: how to adapt Claude Code to your project and your way of working. Modules 7–11.

Automation and integration

Headless mode with claude -p, JSON output, Unix pipelines and CI, remote sessions and asynchronous work. Modules 12–15.

Professional level

The ecosystem and evaluating sources, choosing stacks and architectures, downloadable agentic boilerplates and team workflows. Modules 16–19.

Prerequisites

The only hard requirement is a Claude account with an active plan (Pro, Max or Team). The course always uses the account login: a paid API key is never needed, and never asked for.

These help, but are not mandatory:

• Familiarity with the terminal — you can move between folders, run commands, read an error.
• The basics of git — commit, branch and diff, at least as concepts.
• Node.js installed, for the hands-on examples in some modules.
• macOS or Windows: every command is given in both versions, zsh and PowerShell, with the choice of operating system synchronized across all pages.

No previous experience with AI agents or with Claude Code is needed: we start from zero.

Not a developer?

You can still follow the course. Start with The basics — ten minutes on the terminal, git, file formats and how an AI model reasons — and keep the Glossary open: every technical term in the course is explained there in one sentence, with an analogy. Inside the modules, the more technical steps have a collapsible “Starting from zero” box with the minimum explanation you need to keep going.

Where to start

The basics, before you startTen minutes on the terminal, git and how an AI model reasons. If you are not a developer by trade, start here.

0 · Setup and installationInstall the CLI and the desktop app, log in and open your first project.

2 · Anatomy of the filesIf you have already installed everything, start here: where the configuration really lives.

18 · Agentic boilerplatesThe three downloadable template folders for starting a project in a few minutes.

16 · Ecosystem and sourcesWhere to find reliable skills, MCP servers and plugins — and how to evaluate them.

How to read the modules

Every module always has the same structure, so you know what to expect: estimated time and prerequisites, objectives, key concepts, a step-by-step guided example with the expected output, an exercise with the solution in a collapsible block, common mistakes and troubleshooting, a final cheat sheet and the official sources consulted, with the date they were checked.

And the course runs on two tracks: the main text is written for developers, but every technical term is linked to the Glossary on its first occurrence, and the steps that assume developer knowledge have a collapsible “Starting from zero” box with the minimum explanation needed to follow along. Those who know skip them; those who don’t never get lost.