Getting Started

Getting Started with Claude Code

Claude Code is a CLI tool that puts an AI coding assistant directly in your terminal. You type messages, Claude reads your files, runs commands, and edits code — all with your approval. It’s configurable at every level: you can write custom skills (like slash commands), set rules for what Claude can and can’t touch, and automate actions with hooks.

If you’re coming from a web-based AI chat tool, the big difference is that Claude Code lives in your project. It reads your codebase, respects your .gitignore, and saves memory across sessions.

Already know what you’re looking for? Browse the reference hubs in the sidebar — every env var, flag, setting, and command is cross-referenced there.

Choose Your Learning Path

Not sure where to start? Pick the path that matches what you want to do:

Path A — Configure how Claude behaves

You want to control Claude’s model, API key, verbosity, or session behavior.

CLI flags — flags you pass when starting Claude (--model, --verbose, --print)
Settings — persistent config in .claude/settings.json
Environment variables — env vars that override settings (start with the top-5 list at the top of that page)

Path B — Extend Claude with custom skills

You want to add slash commands, automate workflows, or control what Claude is allowed to do.

Your first skill — create a /slash-command in 5 minutes
Your first permission rule — stop Claude from asking you every time it edits a file
Your first hook — run a script automatically before or after Claude uses a tool
Plugins — package skills, hooks, and MCP servers together for sharing

Path C — Understand what Claude can do in a session

You want to know what slash commands exist, what tools Claude can call, or how memory works.

Slash commands — all built-in /commands
Built-in tools — filesystem, shell, web, and orchestration tools Claude can call
Memory — how Claude remembers things across sessions

First 10 Minutes Checklist

Work through these in order to get oriented:

Install: npm install -g @anthropic/claude-code (or follow the official install guide)
Set your API key: export ANTHROPIC_API_KEY=sk-ant-...
Start a session: claude in any project directory
Type a message — try “what files are in this project?”
Try /help to see all available commands
Try /config to see your current settings
Create a CLAUDE.md in your project root — Claude reads it automatically as project context
Try /memory to see what Claude remembers about your project

Key Concepts

Before diving into the reference docs, these 5 concepts cover 90% of day-to-day use:

Concept	What it is	Where to learn more
CLAUDE.md	A markdown file Claude reads as project context. Put architecture notes, conventions, and “don’ts” here.	Memory overview
Skills	Custom slash commands you write as markdown files. Put them in `.claude/skills/`.	First skill
Permissions	Rules that control which files, commands, and tools Claude can use without asking.	First permission rule
Hooks	Shell scripts or webhooks that run automatically at lifecycle events (before/after tool use, etc.).	First hook
Settings	`.claude/settings.json` — your project’s config file. Checked into git; shared with your team.	Settings overview