Getting Started with Claude Code: A Beginner's Guide

Claude Code is a tool that lets you talk to an AI assistant right inside your computer's terminal. Instead of writing code line by line yourself, you describe what you want in plain English, and Claude finds the right files, makes the changes, and shows you everything before it touches anything. Here's how to get up and running in about ten minutes.

What you need before starting

Just three things:

1. A terminal (Command Prompt or PowerShell on Windows, Terminal on Mac). If you've never opened one, don't worry, you'll only need a few simple commands. 2. A code project on your computer, even a small one. 3. A Claude account. A paid Claude subscription (Pro, Max, Team, or Enterprise) works, and so does a Claude Console account with API credits.

Claude Code also works on the web, as a desktop app, and inside editors like VS Code, but the terminal version is the classic way to use it and what this guide covers.

Step 1: Install it

Open your terminal and run one command.

On Mac or Linux:

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

On Windows (PowerShell):

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

That's it. This version updates itself automatically in the background, so you never have to reinstall.

A quick Windows tip: if you get an error saying `irm` is not recognized, you're in Command Prompt instead of PowerShell. Your prompt shows `PS C:\` when you're in the right place.

Step 2: Log in

In your terminal, type:

claude

The first time you run it, it will ask you to log in. Follow the prompts, and it opens your browser where you sign in with your Claude account. You only do this once; your login is saved after that.

Step 3: Open it inside your project

Claude Code works on whatever folder you start it from. So navigate to your project first, then start it:

cd path/to/your/project
claude

You'll see a prompt where you can start typing. One nice thing: Claude reads your project files on its own as needed. You don't have to copy and paste code into it.

Step 4: Ask it questions

The best first move is to just ask about your own project. Type things like:

what does this project do?

explain the folder structure

where is the main entry point?

Claude will look through your files and explain them in plain language. This alone is incredibly useful when you're handed a project you've never seen before.

Step 5: Let it make a change

Now try an actual edit:

add a hello world function to the main file

Claude will find the right file, show you exactly what it wants to change, and wait for your approval. Nothing gets modified without your permission, so you can experiment freely. If you get comfortable and trust it, you can turn on "accept all" mode for a session so it stops asking every time.

Step 6: Use it for Git

If you use Git, this is where Claude Code really saves time. Instead of remembering Git commands, just say what you want:

what files have I changed?

commit my changes with a descriptive message

create a new branch called feature/quickstart

help me resolve merge conflicts

Step 7: Fix bugs and build features

Describe the problem or the goal the way you'd describe it to a coworker:

there's a bug where users can submit empty forms - fix it

add input validation to the user registration form

Claude locates the relevant code, figures out the context, writes a fix, and runs your tests if you have any. You can also ask it to refactor old code, write unit tests, update your README, or review your changes before you push them.

Commands worth memorizing

You really only need a handful:

| Command | What it does | |---|---| | `claude` | Start a session | | `claude "fix the build error"` | Run a single task and exit | | `claude -c` | Continue your most recent conversation | | `claude -r` | Resume an older conversation | | `/help` | Show all commands (inside a session) | | `/clear` | Clear the conversation history | | `/exit` | Quit |

Tips that make a big difference

Be specific. "Fix the bug" gives Claude little to work with. "Fix the login bug where users see a blank screen after entering wrong credentials" gets you a much better result.

Break big jobs into steps. Instead of one giant request, list them: first create the database table, then the API endpoint, then the page that uses it. Claude works through them in order.

Let Claude look around first. Before asking for changes in unfamiliar code, ask it to analyze the relevant part, like "analyze the database schema." It makes better changes when it understands the surroundings.

Learn the shortcuts. Type `/` to see all commands, press the up arrow for your command history, and use Tab to autocomplete.

If you get stuck

Type `/help` inside Claude Code, or just ask it directly: "how do I...". It can explain its own features. The full documentation lives at code.claude.com/docs, and Anthropic runs a Discord community for tips and support.

The main thing to remember: talk to it like a helpful colleague, not a search engine. Describe what you're trying to achieve, and it will help you get there.