How to use Claude Code: install, first session, CLAUDE.md and permissions

Claude Code is Anthropic’s coding agent. It reads your codebase, edits files, runs commands and works with tools such as Git, and it lives where developers already work: the terminal, VS Code and JetBrains, a desktop app and the web. This tutorial walks through the essentials: installing it, logging in, running your first session, giving it instructions with a CLAUDE.md file and choosing how much it may do without asking. Commands and settings are taken from Anthropic’s documentation as of September 2026.

What you need

• A terminal or command prompt.
• A code project to work on.
• A Claude Pro, Max, Team or Enterprise plan, a Claude Console account, or access through Amazon Bedrock, Google Cloud’s Agent Platform or Microsoft Foundry.

Step 1: Install Claude Code

macOS, Linux and WSL

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

Windows PowerShell

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

Windows Command Prompt

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

Homebrew

brew install --cask claude-code

WinGet

winget install Anthropic.ClaudeCode

Confirm it worked:

claude --version

The command prints a version number followed by “(Claude Code)”. Native installs update themselves; Homebrew and WinGet installs don’t. On native Windows, Anthropic recommends installing Git for Windows so Claude Code can use the Bash tool; without it, Claude Code uses PowerShell.

Step 2: Log in

Start Claude Code with claude. On first use it asks you to log in, and you finish signing in through your browser. If you have set the ANTHROPIC_API_KEY environment variable, Claude Code asks you to approve the key instead. To switch accounts later, type /login inside a session.

Step 3: Start your first session

Open your terminal in a project folder and run:

cd /path/to/your/project
claude

You’ll see the version, the current model and the working directory. Type /help for available commands, or /resume to continue an earlier conversation.

Step 4: Ask your first questions

Start by letting Claude explore. It reads your files as needed, so you don’t have to paste context.

• “what does this project do?”
• “what technologies does this project use?”
• “where is the main entry point?”
• “explain the folder structure”

Step 5: Make your first change

Ask for something small, such as “add a hello world function to the main file”. Claude finds the right file and shows the change. If it asks before editing, select Yes to approve.

Step 6: Use Git conversationally

Claude Code works directly with Git:

• “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 a bug or add a feature

Describe the goal in plain language: “there’s a bug where users can submit empty forms - fix it”. Claude locates the code, works out the context, implements a solution and runs tests if they exist. Be specific. Instead of “fix the bug”, say “fix the login bug where users see a blank screen after entering wrong credentials”. For big jobs, list the steps.

Essential commands

Give Claude instructions with CLAUDE.md

Each session starts with a fresh context window. A CLAUDE.md file carries your instructions across sessions: coding standards, architecture, common commands and review checklists.

• Run /init to generate a starting CLAUDE.md. Claude analyses your codebase and writes build commands, test instructions and conventions it finds. If one already exists, it suggests improvements.
• Run /memory to list and edit the memory files loaded in the session, and to toggle auto memory, notes Claude writes itself from your corrections and preferences.
• Run /context to check which files loaded.
• Keep it short and specific. Claude treats CLAUDE.md as context, not enforced rules; to block an action for certain, use a permission rule or a hook.

If your repository already has an AGENTS.md for other coding agents, Claude Code can read that too.

Choose a permission mode

A permission mode decides what Claude can do without asking. Press Shift+Tab to cycle modes in a session.

On Pro, Max and Team plans, auto mode is the built-in starting mode for interactive terminal sessions; on other plans it is Manual. Run /permissions to see and change the allow and deny rules, which live in settings.json files such as .claude/settings.json in a project and ~/.claude/settings.json for you. Claude Code enforces these rules, not the model, so an instruction in your prompt can’t override them.

Use Claude Code beyond the terminal

• VS Code and Cursor: an extension with inline diffs, @-mentions and plan review.
• JetBrains: a plugin for IntelliJ IDEA, PyCharm, WebStorm and other IDEs; it needs the CLI installed separately.
• Desktop app: macOS and Windows, with visual diffs, side-by-side sessions and scheduled tasks. Open the Code tab. It needs a paid plan.
• Web: claude.ai/code, and the Claude apps for iOS and Android, for long-running tasks with no local setup.
• Slack and CI: mention @Claude in Slack, or run it in GitHub Actions and GitLab CI/CD.

Go further

Once the basics feel natural, Anthropic’s documentation points to MCP servers that connect Claude to tools such as Jira and Slack, skills that package repeatable workflows, hooks that run commands before or after actions, subagents that work in parallel, and routines that run on a schedule, even when your computer is off. You can also pipe output into it, for example tail -200 app.log | claude -p "Slack me if you see any anomalies".

Safety tips

• Work on a branch and review every diff before you merge.
• Start in default or plan mode until you trust what Claude does.
• Keep secrets out of prompts and files Claude can read.
• Use bypassPermissions only in an isolated container or virtual machine.

To see how Claude Code compares with Codex, GitHub Copilot, Cursor and others, read our guide to the best AI coding assistants. For the same approach applied to everyday work, see What is Claude Cowork?, and for plans and models, see Claude Pro vs Max, Claude models explained and the Claude profile.