Introduction to Claude Code Beginner
Claude Code is Anthropic's official command-line interface (CLI) for AI-assisted coding. It brings the power of Claude directly into your terminal, allowing you to edit files, navigate codebases, run commands, and manage git workflows — all through natural language conversation.
What is Claude Code?
Claude Code is an agentic coding tool that lives in your terminal. Unlike browser-based AI assistants, Claude Code operates directly in your development environment. You give it natural language instructions, and it uses a set of tools — reading files, editing code, running shell commands — to accomplish complex coding tasks autonomously.
Think of Claude Code as a pair programmer who sits inside your terminal. You describe what you want, and Claude Code figures out the steps to make it happen — reading the relevant files, making the changes, and verifying the results.
What Claude Code Can Do
Claude Code is designed to handle a wide range of development tasks. Here are the core capabilities:
| Capability | Description | Example |
|---|---|---|
| Edit files | Create, modify, and refactor code across single or multiple files | "Add error handling to the login function" |
| Search codebases | Find files, functions, patterns, and understand project structure | "Where is the authentication middleware defined?" |
| Run commands | Execute shell commands, run tests, build projects | "Run the test suite and fix any failures" |
| Git operations | Create commits, manage branches, generate PR descriptions | "Commit these changes with a descriptive message" |
| Answer questions | Explain code, architectures, and suggest improvements | "How does the caching layer work in this project?" |
| Generate code | Create new files, functions, tests, and documentation | "Write unit tests for the UserService class" |
How It Works
Claude Code follows an agentic loop pattern. When you send a message, here is what happens behind the scenes:
-
You provide a natural language instruction
Type your request in plain English (or any language). For example: "Fix the bug where users can't log in with email addresses containing a plus sign."
-
Claude analyzes and plans
Claude reads your request, considers the context (current directory, recent conversation), and decides which tools to use.
-
Claude takes actions using tools
Claude can read files (to understand context), search the codebase (to find relevant code), edit files (to make changes), and run commands (to test or verify).
-
You review and approve
By default, Claude asks for your approval before making changes. You can review each action and accept or reject it.
-
Iterate as needed
Continue the conversation to refine, adjust, or extend the changes. Claude remembers the full context of your session.
# Start Claude Code in your project directory $ cd my-project $ claude # Claude Code starts and shows a prompt Claude Code > Fix the login bug where emails with + signs are rejected # Claude will then: # 1. Search for login-related files # 2. Read the relevant code # 3. Identify the bug # 4. Propose a fix # 5. Apply the edit (with your approval)
Key Features
Multi-File Editing
Claude Code can make coordinated changes across multiple files in a single session. Need to rename a function and update all its call sites? Add a new API endpoint with route, controller, model, and tests? Claude Code handles it all.
Codebase Understanding
Claude Code can read and understand your entire project structure. It searches for files using glob patterns and grep, reads file contents, and builds a mental model of your architecture — all within the conversation context.
Git Integration
Built-in git support means Claude Code can create well-formatted commits, generate branch names, write PR descriptions, and even review diffs. The /commit slash command makes committing as simple as one word.
Context Persistence
Within a session, Claude Code remembers everything you've discussed. It retains context about files it has read, changes it has made, and your preferences. Use CLAUDE.md files to persist project-level context across sessions.
Permission System
Claude Code has a thoughtful permission system that lets you control how much autonomy it has. Choose from interactive approval, auto-accept for safe operations, or fully autonomous modes.
Supported Platforms
| Platform | Support Level | Notes |
|---|---|---|
| macOS | Full support | Primary development platform. Works natively in Terminal, iTerm2, etc. |
| Linux | Full support | Works on all major distributions. Great for server-side development and SSH sessions. |
| Windows (WSL) | Full support via WSL | Use Windows Subsystem for Linux. Install Node.js inside WSL for the best experience. |
| VS Code | Extension available | Claude Code also has a VS Code extension that embeds the CLI in the editor sidebar. |
Claude Code vs Other AI Coding Tools
There are many AI coding tools available today. Here is how Claude Code compares to the most popular alternatives:
| Feature | Claude Code | GitHub Copilot | Cursor |
|---|---|---|---|
| Interface | Terminal (CLI) | IDE plugin | Full IDE (VS Code fork) |
| Agentic mode | Yes, primary mode | Limited (Copilot Chat) | Yes (Composer) |
| Multi-file edits | Yes | Limited | Yes |
| Shell commands | Yes, native | No | Yes (terminal) |
| Git integration | Deep (commits, PRs, diffs) | Basic | Basic |
| Works over SSH | Yes | With remote extension | With remote SSH |
| No IDE required | Yes | No | No |
| Model | Claude (Sonnet/Opus) | GPT-4 / Claude | Multiple (GPT-4, Claude, etc.) |
Ready to Get Started?
Now that you understand what Claude Code is and what it can do, head to the next lesson to install it on your system and start using it.
Next: Installation →
Lilly Tech Systems