Welcome to the world of AI agents that actually execute.
For official documentation, see the Anthropic Claude Code docs. MCP specifications are available at modelcontextprotocol.io.
Claude Code lets you control Anthropic’s Claude AI directly from your terminal. Unlike traditional chat UIs, this Claude Code tutorial will show you how file operations, browser automation, API integrations, and scheduled execution all happen inside the terminal. In other words, you are not just asking AI for help — you are working alongside it.
This is the first article in a series where I share the complete automation environment I have built with Claude Code. Moreover, each article covers a specific aspect of the setup, from project structure to desktop automation.
What Is Claude Code?

Claude Code is a CLI-based AI coding agent developed by Anthropic. Think of it as having an AI that lives inside your terminal. For example, when you ask it to fix a bug, it reads the relevant files, edits the code, runs tests, and even handles browser operations if needed.
However, it is fundamentally different from a simple code completion tool. Claude Code can read and write files, execute shell commands, operate websites, and send emails. Essentially, it is like having an additional assistant sitting at your computer, ready to perform any task you describe.
How Claude Code Differs from ChatGPT and Copilot
You might wonder, “Why not just use ChatGPT or GitHub Copilot?” However, Claude Code offers decisive advantages that set it apart from both.
First, there is execution capability. While ChatGPT merely suggests code, Claude Code actually creates and edits files, then executes commands. Furthermore, whereas Copilot is limited to in-editor completions, Claude Code’s reach extends to the terminal, browser, and desktop applications.
In addition, Claude Code supports MCP (Model Context Protocol) for extensibility. You can connect external tools like Gmail, Google Calendar, Chrome, and Windows desktop as plugins. As a result, Claude Code can automate not just coding tasks, but virtually any daily workflow.
What You Can Actually Do with Claude Code
Here is what I automate with Claude Code on a daily basis:
- Browser automation — Purchasing on Amazon, making restaurant reservations, searching for flights
- Desktop control — Sending LINE messages via the PC app, automating arbitrary applications
- Scheduled tasks — Monthly invoice generation and submission, daily news digest delivery
- Financial management — Building and running a Streamlit dashboard integrated with the freee accounting API
- Persistent memory — Retaining user preferences and project context across conversation sessions
For instance, saying “Buy tissues on Amazon” triggers an automated workflow from product search to cart addition, pausing just before order confirmation. Meanwhile, saying “Message my mother on LINE” makes Claude Code operate the desktop LINE app to send the message directly.
📊 Claude Code Overview
About This Claude Code Tutorial Series
This Claude Code tutorial series covers my complete automation setup across eight articles. Specifically, each installment covers the following:
- Introduction (this article) — Overview and installation guide
- Project Structure — CLAUDE.md and directory design best practices
- MCP Servers — Connecting external tools via custom MCP servers
- Skills — One-command custom automations for purchases, reservations, and messaging
- Memory — Building a persistent context system across conversations
- Scheduled Execution — Cron-like automated recurring tasks
- App Development — Pair programming with AI to build a finance dashboard
- Desktop Automation — Controlling Windows and browsers with AI
You can read the articles in any order. However, if this is your first time using Claude Code, starting with this introduction is recommended.
Prerequisites and Installation
To get started with this Claude Code tutorial, you need the following:
- Anthropic account — Either the Max plan (monthly subscription) or API billing
- Node.js 18+ — Required for npm installation
- Terminal environment — Works on Windows, Mac, and Linux
Once you have the prerequisites, installation takes just one command:
npm install -g @anthropic-ai/claude-code
After installation, navigate to any project directory and run:
claude
On first launch, you will be prompted to authenticate your Anthropic account. Simply log in through your browser, and you are ready to go.
What to Try First
Once you have Claude Code installed, try these simple commands to see it in action:
Explain this project structure— Claude Code reads and interprets your directory layoutCreate a README— Auto-generates documentation from your projectFix this bug— Paste an error message and Claude Code identifies the cause and applies the fix
The key point is that Claude Code does not just suggest answers — it actually edits your files. You review and approve each change before it is applied, so there is no risk of unintended modifications.
🔍 Which Article Should You Read First?
“I want to try Claude Code” → Start with this Introduction
“I want to organize my project” → Project Structure
“I want to connect external tools” → MCP Servers
“I want to automate daily tasks” → Custom Skills
“I want AI to remember context” → Memory System
“I want unattended recurring jobs” → Scheduled Tasks
“I want to build an app with AI” → App Development
“I want to control browser and desktop” → Desktop Automation
Next Up
The next article covers Project Structure. You will learn how to write the CLAUDE.md configuration file and organize your skills, schedules, and MCP servers for maximum efficiency. Additionally, we will discuss best practices for long-term project maintenance.
