What is Claude Code?

You are about to use something that did not exist two years ago.

Claude Code is an AI that lives in your terminal. You open it, describe what you want in plain English, and it writes the code, creates the files, runs the commands, and builds the thing. You do not need to know how to code. You do not need to understand programming languages, frameworks, or build tools. You describe. It executes.

This is not a chatbot. Chatbots answer questions. Claude Code does things. It reads your files, understands your project, creates new files, edits existing ones, runs terminal commands, and checks its own work. When you tell it “build me a personal website,” it does not explain how to build one. It builds one.

This is not autocomplete. Autocomplete suggests the next few characters as you type code. Claude Code writes entire projects from a conversation. You never touch a code editor unless you want to.

Think of it this way: you are hiring a developer who works at your direction. You say what you want. They figure out how to build it, ask clarifying questions if needed, and deliver working code. The difference is that this developer works in seconds, costs nothing extra on your current plan, and never gets tired.

By the end of this lesson, you will have Claude Code installed and running on your machine. You will have had your first conversation with it. And you will have watched it create something real on your computer.

Let’s start.

You need two things on your computer: Node.js and Claude Code itself. This takes about five minutes.

Step 1: Install Node.js

Node.js is the engine that runs Claude Code. You do not need to understand what it does. You just need it installed.

1. Go to nodejs.org
2. Download the LTS version (the green button)
3. Run the installer. Accept all the defaults.

To verify it worked, open your terminal:

• Mac: Open the Terminal app (search “Terminal” in Spotlight)
• Windows: Open PowerShell (search “PowerShell” in Start)
• Linux: You already know where your terminal is

Type this and press Enter:

node --version

You should see a version number like v22.12.0. The exact number does not matter. If you see an error like command not found, close your terminal, reopen it, and try again. The installer sometimes needs a fresh terminal to take effect.

Step 2: Install Claude Code

Now install Claude Code itself. In your terminal, type:

npm install -g @anthropic-ai/claude-code

This downloads Claude Code and makes it available everywhere on your machine. It takes 30-60 seconds.

Step 3: Verify it works

Type:

claude

You should see a welcome screen. Claude Code is running. It will ask you to log in with your Anthropic account the first time. Follow the prompts.

Troubleshooting

“command not found” after installing Node.js: Close your terminal completely and open a new one. If that does not work, restart your computer.

“permission denied” on Mac/Linux: Run the install with sudo:

sudo npm install -g @anthropic-ai/claude-code

“npm is not recognized” on Windows: Make sure you checked “Add to PATH” during the Node.js installation. If not, uninstall Node.js and reinstall, checking that box.

Already have Node.js but it is old: Claude Code needs Node.js 18 or newer. Run node --version to check. If it is below 18, download the latest LTS from nodejs.org.

Claude Code is running. Time to use it.

Open Claude Code in a folder

Open your terminal and navigate to any folder. Your home directory works fine for now:

cd ~
claude

You are now inside Claude Code. You will see a prompt waiting for your input. Everything you type here is a conversation with an AI that can see and modify files on your computer.

Ask it something

Type this:

What files are in this folder?

Watch what happens. Claude Code will use its tools to explore your file system. It will list what it finds. This is not a guess or a memory. It is reading your actual file system in real time.

Create something

Now type this:

Create a file called hello.txt that says "I built this with AI"

Claude Code will ask your permission before creating the file. You will see a prompt like:

Claude wants to create file: hello.txt
Allow? [y/n]

Type y and press Enter.

Now check. Open your file explorer or run cat hello.txt in a new terminal. The file is there. The text is inside it. Claude Code created a real file on your real computer.

This is the core loop. You describe what you want. Claude Code asks permission. You approve. It executes. Everything it does is visible and reversible. You are always in control.

The permission system

Claude Code never silently modifies your computer. Every file creation, edit, or command execution requires your approval. You will see exactly what it wants to do before it does it.

As you build trust, you can allow certain operations automatically. But for now, approving each action is the right approach. It lets you see exactly what Claude Code does and how it thinks.

End your session

When you are done, type:

/exit

Or press Ctrl+C. Your conversation is saved. Next time you open Claude Code in the same folder, it remembers the context.

You have seen Claude Code in action. Before you start building real projects in the next lesson, it helps to understand how it works under the hood. Not the technical details, but the mental model.

It reads everything

When you open Claude Code in a folder, it can read every file in that folder and its subfolders. It understands project structure, configuration files, code in dozens of programming languages, and plain text. When you ask it to do something, it uses this context to make informed decisions.

This is why working inside a project folder matters. Claude Code in an empty folder has no context. Claude Code inside your project knows what already exists and builds on top of it.

It uses tools

Claude Code does not just generate text. It uses specific tools to interact with your computer:

• Read: Opens and reads files
• Write: Creates new files
• Edit: Makes precise changes to existing files
• Bash: Runs terminal commands (install packages, start servers, run tests)
• Grep: Searches through files for specific text
• Glob: Finds files matching a pattern

When you say “add a contact form to my website,” Claude Code reads your existing HTML, decides where the form should go, uses Edit to insert it, and might use Bash to install a form-handling library. It chains these tools together automatically.

It plans before acting

Watch Claude Code work on a non-trivial task and you will notice it thinks before acting. It reads relevant files first, considers the approach, then executes step by step. If something goes wrong, it reads the error, adjusts, and tries again.

You can see this thinking in real time. Claude Code shows you what it is doing and why. This transparency is a feature, not a bug. It means you can course-correct early instead of discovering problems after the fact.

You direct, it executes

The most important mental model: you are the architect, Claude Code is the builder.

You decide what to build. Claude Code figures out how to build it. You approve each action. You can interrupt at any point, change direction, ask questions, or tell it to try a different approach.

Good prompts are specific about the outcome, not the implementation:

• “Build me a personal website with my name, bio, and links to my social accounts”
• “Add a dark mode toggle that remembers the user’s preference”
• “Make the contact form send emails to hello@mysite.com”

You do not need to say “use React” or “add a CSS class” or “create a function.” Claude Code makes those decisions based on your project’s existing patterns. If you have opinions about implementation, you can express them. But you do not have to.

What’s next

You now have Claude Code installed, you have had your first conversation, and you understand the mental model. In the next lesson, you will use all of this to build your first real project: a personal landing page, from scratch, deployed to the internet.