Skip to content

Claude Code — Agentic Coding CLI by Anthropic

Claude Code is Anthropic’s terminal-based AI coding agent. Unlike IDE plugins that add AI to your existing editor, Claude Code operates in your terminal — reading your files, running shell commands, editing code, and committing changes, all driven by natural language instructions.

You run claude in your terminal, describe what you want, and Claude Code:

  1. Reads relevant files in your codebase
  2. Plans what needs to change
  3. Makes edits across multiple files
  4. Runs tests or build commands to verify changes
  5. Fixes errors if tests fail
  6. Commits and describes the changes

All while keeping you in control — you can see every action and approve or reject changes.

Terminal window
# Install via npm (requires Node.js 18+)
npm install -g @anthropic-ai/claude-code
# Verify installation
claude --version
# Start a session in your project directory
cd my-project
claude
CapabilityExample
Read codebaseUnderstands your entire project structure
Write + edit filesCreates and modifies files
Run commandsExecutes shell commands, tests, builds
Git operationsStages files, creates commits, reads diffs
Web searchLooks up documentation when needed
Multi-turn contextRemembers the full conversation and file state
Terminal window
claude "The login form is not validating the email format. Fix it and add a test."

Claude Code will:

  • Find the login form component
  • Identify the missing validation
  • Add the validation logic
  • Write a unit test for it
  • Run the test suite to confirm it passes
Terminal window
claude "Add a dark mode toggle to the settings page. Persist the preference in localStorage."
Terminal window
claude "Refactor the UserService class to use dependency injection and write tests for each method."
Terminal window
claude "Explain how authentication works in this project and draw a flow diagram."
CommandDescription
/helpShow available commands
/clearClear conversation history
/compactCompact the conversation to save context
/configOpen configuration settings
/memoryView or edit persistent memory
/reviewReview recent changes

Create a CLAUDE.md file at the root of your project to give Claude Code persistent instructions:

# Project: MyApp
## Stack
- Frontend: React + TypeScript
- Backend: Node.js + Express
- Database: PostgreSQL
## Conventions
- Use functional components (no class components)
- All functions must have JSDoc comments
- Tests go in __tests__ folder next to the source file
- Run `npm test` to verify changes
## Rules
- Never commit .env files
- Always run `npm run lint` before finishing

Claude Code Cloud allows Claude Code to run in a secure, sandboxed cloud environment — useful for:

  • Running long tasks without keeping your terminal open
  • Executing in CI/CD pipelines
  • Sharing sessions with team members
FeatureClaude CodeCursorWindsurf
InterfaceTerminalIDE (GUI)IDE (GUI)
Multi-file editsYesYesYes
Runs shell commandsYes (core feature)Yes (agent mode)Yes (Cascade)
Git integrationYes (native)YesYes
ModelClaude onlyClaude, GPT-4o, GeminiClaude, GPT-4o
Works in any editorYes (terminal)VS Code fork onlyVS Code fork only
Learning curveMediumLowLow
Best forAutonomous tasksVisual AI editingFast AI editing

Claude Code requires a Claude Pro plan ($20/month) or API access. Usage is billed against your Anthropic API usage in API mode.