Skip to main content

Your first day in Claude Code

Goal: In about 15 minutes, you’ll open Claude Code, sign in, and complete your first AI-assisted change. This guide covers only what you need on day one; links to deeper material appear at the end.


Step 1: Open Claude Code (2 minutes)

Recommended: Download Claude Desktop, sign in with your paid plan, and open the ā€œCodeā€ tab. Claude Desktop includes Claude Code, so there’s nothing else to install. Follow the desktop quickstart in the Claude Code docs for your first session, then skip to Step 3.

To use the terminal instead, pick the command that matches your machine and paste it into a terminal.

Your setup

Command

macOS / Linux / WSL

Windows (PowerShell)

Homebrew

brew install --cask claude-code

npm (Node 18+)

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

Tip: Do not run the npm install with sudo, as this causes file-permission problems later.

Verify the installation:

claude --version

Step 2: Sign in (2 minutes)

If you’re using Claude Desktop, you’re already signed in and Claude Code uses the same account. Skip to Step 3.

In the terminal, how you authenticate depends on how your organization has provisioned access. Pick one of the following:

If your organization uses…

Do this

A paid Claude plan

Run claude, then /login and approve the request in your browser (including SSO if prompted). This is the most common path.

An Anthropic Console API key

Run export ANTHROPIC_API_KEY=sk-ant-…, then run claude.

Amazon Bedrock

Run export CLAUDE_CODE_USE_BEDROCK=1, configure your AWS credentials, then run claude.

Google Vertex AI

Run export CLAUDE_CODE_USE_VERTEX=1, configure your GCP credentials, then run claude.

If you are unsure which method applies, check with your administrator or try /login first, which is the default for most organizations.


Step 3: Open a project (1 minute)

In Claude Desktop, start a new session in the ā€œCodeā€ tab and choose the folder for the repository you want to work on. In the terminal, start Claude Code from inside the repository:

cd path/to/your/project
claude

Claude Code reads your files on demand, so there’s nothing to upload or attach.


Step 4: Understand the one safety rule (1 minute)

Claude will always ask before changing a file or running a command. When it proposes a change, you will see a diff and three choices:

  • Yes — apply this single change.

  • Yes, and don't ask again for edits — auto-approve file edits for the remainder of the session.

  • No — reject the change and explain why.

You can press Shift+Tab at any time to cycle between modes (Plan → Accept Edits → default). If a change produces an unexpected result, your git history remains untouched until you commit, so git checkout will undo it.


Step 5: Five things to try first

Copy and paste these into the prompt one at a time. Each example demonstrates a different core capability.

1. Get oriented

Give me a 5-bullet summary of what this codebase does and where the entry point is.

Why this is useful: It shows how Claude explores files on its own, without being pointed at specific paths.

2. Find something

Where is user authentication handled? Show me the file and the key function.

Why this is useful: It is often faster than grep when you do not know the exact symbol or filename.

3. Make a safe edit

Add a docstring to the function <name> in <file>. Keep it to 2 lines.

Why this is useful: It lets you practice reviewing and approving a diff on a low-risk change.

4. Fix something real

This test is failing: <paste the error>. Find the cause and fix it.

Why this is useful: This is the primary workflow—you describe the symptom, and Claude investigates and proposes a fix.

5. Let it handle git

Stage my changes and write a commit message that follows our existing style.

Why this is useful: Claude can run git on your behalf (with approval) and will match your repository's existing commit conventions.


Step 6: Before you finish for the day

Run this once per project:

/init

This generates a CLAUDE.md file at your project root that captures your codebase's conventions. Claude reads it automatically at the start of every session, so future responses will already be tailored to your project. It is the single highest-value setup step you can take.


Next steps

You have now opened Claude Code, signed in, made an edit, and set up project memory. When you are ready to go further, the official documentation covers memory and CLAUDE.md, built-in commands, and interactive-mode shortcuts in depth.

If you ran into problems during install or login, see the troubleshooting guide.

Did this answer your question?