AI Reference
Claude Code
Anthropic's terminal-based AI coding agent. The most-used AI coding tool in 2026.
Last updated: October 2026
Claude Code is Anthropic's terminal-based AI coding agent. Describe a task in plain English, and it reads your codebase, edits files, runs tests, and reports back — autonomously.
Full definition: Claude Code runs in your terminal, operates on your local files, and can complete multi-file tasks without step-by-step instructions. Unlike autocomplete tools, it plans and executes entire workflows.
📊 By the Numbers (2026)
| Metric | Value |
|---|---|
| Workplace usage (JetBrains 2026) | 39% — #1 AI coding tool |
| Growth since early 2025 | 3% → 39% (13×) |
| Annual run-rate revenue | $2.5B+ |
| Awareness | 79% |
Install
# macOS, Linux, WSL
curl -fsSL https://claude.ai/install.sh | bash
# Windows PowerShell
irm https://claude.ai/install.ps1 | iex
# Verify
claude --versionHomebrew and WinGet installs are also available. Native installs auto-update.
Authenticate
cd /path/to/your/project
claudeOpens browser for login. Works with Claude Pro/Max subscription or Anthropic Console API key.
Subscription required. Pro is $20/month, Max starts at $100/month.
Core Commands
claude # Start interactive session
claude -p "fix the lint errors" # One-shot prompt, no session
claude --continue # Resume last session
claude --resume # Pick a session from historySlash Commands
/help # Show all slash commands
/clear # Reset context
/compact # Compress context to save tokens
/init # Generate CLAUDE.md for your project
/model # Switch Claude model
/review # Request a code reviewCLAUDE.md — Project Context
Add persistent instructions to your project root:
# Project: My App
## Stack
- React 19 + Vite + Tailwind CSS
- TypeScript strict mode
## Conventions
- Functional components only
- Named exports (no default exports)
- Tests with Vitest
## Never
- Commit directly to main
- Modify files in /legacy without approvalClaude Code reads CLAUDE.md automatically at session start.
Real-World Workflow
# Give a bounded task
claude> Create an Express server with three endpoints:
GET /health, POST /users, GET /users.
Add input validation and write tests for all three.Claude Code will create files, install dependencies, write tests, run them, and report results.
🎯 Claude Code vs Cursor vs Copilot
| Feature | Claude Code | Cursor / Copilot |
|---|---|---|
| Interface | Terminal CLI | GUI IDE / extension |
| Multi-file edits | ✅ Native | ✅ Via agent mode |
| Autonomous runs | ✅ Long sessions | ⚠️ Bounded |
| Inline autocomplete | ❌ None | ✅ Best-in-class |
⚠️ Common Mistakes
- Working without version control. Always work in a Git repo. Review diffs before committing.
- Skipping CLAUDE.md. Project context dramatically improves output quality.
- Running long sessions without /clear. Context bloat degrades output.
- Trusting output blindly. Always review generated code before merging.
- Not handling permissions. Claude Code asks before editing. Read what it wants.
❓ Frequently Asked Questions
What is Claude Code?
Claude Code is Anthropic's terminal-based AI coding agent. You describe a task in plain English, and it reads your codebase, edits files, runs tests, and commits results autonomously.
Is Claude Code free?
No. It requires a Claude Pro subscription at $20/month or Max at $100/month, or API access via Anthropic Console.
What are the main Claude Code commands?
claude starts an interactive session, claude -p takes a one-shot prompt, /clear resets context, and /help shows all slash commands. Subagents and hooks extend capabilities.
How do I add project rules to Claude Code?
Create a CLAUDE.md file in your project root. Claude Code reads it automatically at session start and follows the conventions you define.
Does Claude Code work with any language?
Yes. Claude Code is language-agnostic. It works with any codebase — JavaScript, Python, Go, Rust, or mixed repos.