What is Claude Code?

Claude Code is an agentic coding tool built by Anthropic. It reads your codebase, edits files, runs commands, and works across your development tools — all from a conversation in plain English. It is the fastest way to get a capable AI collaborator working inside your actual project, not just answering questions about code in a chat window.

An agentic coding tool

Most AI chat tools answer questions. Claude Code takes actions. It reads the files in your project, proposes changes, runs tests, and commits to git — then asks for your approval before anything is saved. Think of it less like a search engine and more like a junior developer who can see your whole codebase and do real work, but always checks in before touching anything.

Agentic means it acts, not just answers

Claude Code reads your project files as needed without you manually copying and pasting code into a chat. It proposes changes and always asks for permission before modifying files.

What you can do with it

Once Claude Code is running in your project directory, you can hand it real tasks in plain language. Here are some common examples straight from the docs:

Write tests for untested code, run them, and fix failures
Build a new feature across multiple files from a single description
Fix a bug by pasting the error message — it traces the root cause and patches it
Create git commits and open pull requests with descriptive messages
Refactor, update documentation, or resolve merge conflicts
Pipe in log output and ask for an anomaly analysis
Review changed files for security issues before a merge

bash

# Run a one-off task without opening an interactive session
claude "write tests for the auth module, run them, and fix any failures"

# Pipe shell output directly into Claude
tail -200 app.log | claude -p "let me know if you see any anomalies"

# Review changed files for security issues
git diff main --name-only | claude -p "review these changed files for security issues"

Where it runs

Claude Code is not a single app — it is an engine available across several surfaces. Your project instructions (CLAUDE.md), settings, and connected tools work the same way everywhere:

Terminal CLI — the full-featured experience; edit files, run commands, pipe data
VS Code and JetBrains extensions — inline diffs and context sharing inside your editor
Desktop app — visual diff review, multiple sessions side by side, scheduled tasks
Web (claude.ai/code) — no local setup required; start tasks from any browser
GitHub Actions / GitLab CI/CD — automated code review and issue triage in CI
Slack — mention @Claude with a bug report and receive a pull request back

How this guide is organized

This guide takes you from zero to productive in a structured order. You will start with installation and your first session, then move into the features that make Claude Code powerful for daily work: memory and project instructions, slash commands, MCP tool connections, hooks for automation, and running sub-agents for bigger tasks. Each lesson is focused and practical — read them in order or jump to what you need.

Key takeaways

Claude Code is an agentic tool that reads your whole codebase and takes real actions — writing code, running commands, committing to git — while always asking before it changes anything. It runs in your terminal, your editor, a desktop app, and the browser, and the same settings follow you everywhere. The next lessons cover installation and your first session. For the complete feature overview, see the official docs at code.claude.com/docs/en/overview.