What Claude Code is, and installing it

What you’ll be able to do

By the end of this lesson Claude Code will be installed and signed in, and you’ll have asked it a first question about a real folder on your machine, and seen how it worked out the answer.

Why this matters: chat assistant vs. agent

If you’ve used an AI chat assistant in a browser, you know the routine: you copy code into the chat, it suggests changes, you paste them back and run them yourself. The assistant only knows what you show it, and it can’t check whether its suggestion actually works.

Claude Code works inside your project instead. It can list and read your files, search the codebase, run commands such as your tests or build, and edit files. It works in a loop: it takes an action, looks at the result, and decides the next step, until the task is done. Actions that change something, such as editing a file or running a command, go through a permission system, so you stay in charge of what actually happens on your machine.

That changes the kind of request you make. Instead of asking “how would I write a validation function?”, you can ask “add input validation to the signup form and make sure the tests pass”, and Claude does the reading, writing and checking itself.

Step 1: Check the prerequisites

You need:

• A terminal: Terminal on macOS, Windows Terminal or PowerShell on Windows, or any shell on Linux.
• A Claude account with Claude Code access: a Pro, Max, Ultra, Team or Enterprise plan, or a Claude Console account with API billing. The free claude.ai plan doesn’t include Claude Code.
• Git. Claude Code works without it, but every lesson in this course assumes it, because git is your permanent safety net for code changes. On Windows, Git for Windows is optional: without it, Claude Code runs commands in PowerShell instead of Bash.
• Node.js 22 or later, for the practice exercises, which build small JavaScript projects. Check with node --version, or install the LTS version from nodejs.org.

Step 2: Install

Use the native installer for your platform. Pick the command that matches the shell you’re in.

macOS / Linux / WSL

curl -fsSL https://claude.ai/install.sh | bash

Windows (PowerShell): the prompt starts with PS C:\

irm https://claude.ai/install.ps1 | iex

Windows (CMD): the prompt starts with C:\ without PS

curl -fsSL https://claude.ai/install.cmd -o install.cmd && install.cmd && del install.cmd

The installer shows little or no progress while it downloads. Wait for it to finish.

The native installer keeps itself up to date in the background, which matters because Claude Code changes often. Package managers work too, but then you update yourself: brew install --cask claude-code (macOS), winget install Anthropic.ClaudeCode (Windows), or npm install -g @anthropic-ai/claude-code (needs Node.js 22 or later).

Step 3: Start it and sign in

Open a new terminal window. The installer added Claude Code to your PATH (the list of folders your shell searches for commands), but terminals that were already open don’t see that change. Then run:

claude --version
claude

claude --version should print a version number, such as 2.1.318 (Claude Code). Running claude on its own starts an interactive session. The first time, it walks you through signing in with your Claude account in your browser. When that’s done, you’ll see a prompt waiting for your input.

Step 4: Ask your first question

Quit for now by typing /exit (or pressing Ctrl+C twice). Then cd into any project you have, even a small one, and start Claude Code from that folder:

cd ~/code/some-project
claude

The first time you start Claude Code in a folder, it asks whether you trust that folder. Trusting it allows the project’s own Claude Code configuration (which you’ll meet in later lessons) to take effect. Say yes for your own projects; be careful with code you’ve downloaded from someone you don’t know.

Now ask:

What does this project do? Give me a 5-bullet summary and point out the main entry point.

Watch what happens before the answer appears. Claude doesn’t guess: you’ll see it list folders and read files, such as the README and package or config files, and only then write its summary. This “look first, then act” behaviour is the heart of how Claude Code works.

Verify it worked

• claude --version prints a version number
• You signed in and reached the interactive prompt
• You started Claude Code from a project’s root folder and accepted the trust prompt
• Claude summarised a real project, and you saw it read files to do so

Common problems

Symptom	What to do
command not found or “not recognized”	Open a new terminal. If it still fails, add the install folder the installer printed to your PATH.
The token '&&' is not a valid statement separator	You’re in PowerShell, not CMD. Use the PowerShell command.
'irm' is not recognized	You’re in CMD, not PowerShell. Use the CMD command.
Browser sign-in doesn’t come back to the terminal	Copy the URL shown in the terminal into your browser by hand.
Corporate network or proxy errors	Set the HTTPS_PROXY environment variable before starting claude, or ask your IT team.
Anything else	Run claude doctor in your shell. It checks your installation and settings and suggests fixes. Inside a session, /doctor runs a fuller checkup.

Next: what actually happens when you give Claude a task, and how to stay in control of it.