The One-File Wonder

Step 1: Install Claude Code

macOS

• Open Terminal (Cmd+Space, type “Terminal”)
• Paste this and press Enter:

  curl -fsSL https://claude.ai/install.sh | bash

Already use Homebrew? You can also run brew install claude-code instead.

Windows (WSL) — Recommended for Windows

WSL (Windows Subsystem for Linux) gives you a Linux environment inside Windows. Many developer tools work better here, and we’ll need it in later lessons.

1. Open PowerShell (search for it in Start menu)
2. Type this and press Enter:

   wsl --install

3. Restart your computer when it finishes
4. After restart, open the Ubuntu app from your Start menu
5. It will ask you to create a username and password — pick something simple, you’ll need these later
6. Once you see a prompt, paste this and press Enter:

   curl -fsSL https://claude.ai/install.sh | bash

Note: Throughout this course, when we say “open your terminal,” Windows WSL users should open the Ubuntu app, not PowerShell.

Windows (Native)

Heads up: WSL (the option above) is recommended. Some commands later in the course work better in a Linux environment. Native Windows works for now, but you may need to switch later.

• Open PowerShell (search for it in Start menu)
• Paste this and press Enter:

  irm https://claude.ai/install.ps1 | iex

Linux

• Open your terminal
• Paste this and press Enter:

  curl -fsSL https://claude.ai/install.sh | bash

Step 2: Get a Claude subscription

Claude Code runs on a paid Claude plan — it uses your subscription, no API key needed.

Two options:

• Claude Pro ($20/month) — good enough to start learning. From experience: gives about 1.5 hours of active coding, then a ~3 hour cooldown until the limit resets. Fine for working through lessons at a relaxed pace.
• Claude Max ($100/month) — for full-time work. Practically unlimited — you’ll likely never hit the limit.

Our recommendation: Start with Pro. If you find yourself constantly hitting the limit and waiting, upgrade to Max.

Sign up at claude.ai if you don’t already have an account.

Step 3: Learn a few terminal basics

A terminal is a text-based way to talk to your computer — you type a command, press Enter, and see the result.

Here are the commands you’ll use most often:

What you want to do	macOS / Linux / WSL	Windows Native (PowerShell)
See where you are	pwd	cd (no arguments)
List files here	ls	dir
Go into a folder	cd folder-name	cd folder-name
Go up one folder	cd ..	cd ..
Create a folder	mkdir folder-name	mkdir folder-name
Create an empty file	touch filename	New-Item filename
Clear the screen	clear	cls

Note: WSL users use the macOS / Linux / WSL column — WSL is a Linux environment.

Step 4: Create your project folder and start Claude Code

Don’t just create your project in the default terminal location — put it somewhere you can see it, like your Desktop. That way you can find and manage the files using your normal file explorer.

macOS

cd ~/Desktop
mkdir trip-planner
cd trip-planner

Windows (WSL)

WSL can access your Windows files through /mnt/c/. We’ll put your project on the Desktop so you can see it in File Explorer.

1. Type cd /mnt/c/Users and press Enter
2. Type ls — this shows all user folders. Find yours (it’s usually your name or the name you picked when setting up Windows).
3. Type cd YOUR-FOLDER/Desktop (replace YOUR-FOLDER with what you saw in the list)
4. Then:

   mkdir trip-planner && cd trip-planner

Windows (Native)

cd $HOME\Desktop
mkdir trip-planner
cd trip-planner

($HOME is a shortcut PowerShell understands — it points to your user folder automatically.)

Linux

cd ~/Desktop
mkdir trip-planner
cd trip-planner

Start Claude Code

Now, in your project folder, type:

claude

The first time you run it, it will open your browser to log in with your Claude account. Sign in and come back to the terminal.

You’re in. You should see a prompt where you can talk to Claude directly inside your project folder.

Quick tip: What am I looking at?

Claude Code runs in your terminal. When you type a message, Claude reads your project files, thinks about what you asked, and makes changes directly. You’ll see it create files, edit code, and run commands — all in real time. If you have a code editor open (like VS Code or Cursor), you’ll see the files appear and update as Claude works.

Option A: Google Gemini (Recommended for your app)

1. Go to aistudio.google.com
2. Sign in with your Google account
3. Click “Get API Key” in the left sidebar
4. Click “Create API key”
5. Copy the key — it looks something like AIzaSyD...long-string...
6. Save it somewhere safe (a notes app, NOT in your code — we’ll learn why in Lesson 3)

Pricing: Gemini Flash is free for up to 15 requests per minute, 1 million tokens per day. You won’t hit this limit while learning. Paid usage is ~$0.075 per million input tokens — far cheaper than alternatives.

Option B: OpenAI (GPT)

1. Go to platform.openai.com
2. Create an account (email + phone verification)
3. Go to API Keys in the left sidebar
4. Click “Create new secret key”
5. Copy and save the key

Pricing: New accounts get $5 in free credits. GPT-4o costs ~$2.50 per million input tokens. Good for learning, but you’ll burn through free credits faster than Gemini.

Option C: Anthropic (Claude API)

1. Go to console.anthropic.com
2. Create an account (or sign in if you have one)
3. Go to API Keys → Create Key
4. Copy and save the key

Pricing: Claude Sonnet costs ~$3 per million input tokens. Note that the Claude API uses separate billing from your Claude Pro/Max subscription — it charges per use.

Which one should I pick?

If you don’t have a strong preference, go with Gemini Flash. It’s the cheapest, has the most generous free tier, and — this is actually useful for the course — it’s sometimes a bit slow on complex requests. That slowness becomes the perfect teaching moment in Lesson 16 when we learn about async operations.

The Prompt

⚠️ DON’T run this prompt. Read it as an example. It will work — it will build the app — but it puts your API key directly in the code, which is a security risk. We’ll build this app properly in the next section. For now, just read and understand what the prompt is doing.

Build me a trip planner app.

I want to type in a destination and number of days, and it uses the Google Gemini API
to plan my trip day by day.

For each day, show me a morning activity, an afternoon activity, an evening activity,
and a restaurant recommendation.

I want this to be a simple app I can open in my browser. No complicated setup —
just one file I can open directly.

Here's my Gemini API key: [PASTE YOUR KEY HERE]

Make it look clean and easy to read.

If you already ran this prompt: Go back to your AI provider (Gemini / OpenAI / Claude) and DELETE the API key you just used. You pasted it directly into the code, which means it’s sitting in plain text on your computer. Anyone who gets access to that file can use your key and run up your bill. Delete it now, and create a fresh one when we set up proper key management in Lesson 3.

Before moving to Lesson 2, try these in Claude Code:

1. Change the destination — open the app and try different cities, countries, or even “a road trip through Japan.” See how the AI adapts.

2. Ask Claude Code to change things — try saying:

   • “Add estimated costs for each activity”
   • “Make it suggest local street food, not tourist restaurants”
   • “Add a packing list based on the destination and season”

   Watch Claude Code edit the file. Refresh your browser. See the changes.

3. Break it and fix it — tell Claude Code “delete something random from the code.” Then look at the broken app in your browser. Then tell Claude Code “the app is broken, can you fix it?” This is how real vibe coding works — break, ask, fix, repeat.

4. Ask Claude Code to explain it — type “explain what each part of this code does in simple terms.” Claude Code can read the project and walk you through it.

The trip planner is our course project, but the concepts apply to anything. Here’s the same prompt with blanks — try it in Claude Code:

Build me a [DESCRIBE YOUR APP — e.g., "meal planner", "quiz maker", "story generator"].

I want to [WHAT YOU DO — e.g., "type in my dietary restrictions and how many days"]
and it uses the [Gemini / OpenAI / Claude] API to
[WHAT THE AI DOES — e.g., "generate a week of recipes with ingredients and instructions"].

Show me [WHAT YOU WANT TO SEE — e.g., "each day's meals with a shopping list at the bottom"].

I want this to be a simple app I can open in my browser. No complicated setup.

Here's my API key: [YOUR KEY]

Make it look clean and easy to read.

Ideas to try:

• An AI meal planner — type in dietary restrictions, get a week of recipes
• An AI study guide — paste a topic, get flashcards and quiz questions
• An AI cover letter writer — paste a job listing, get a tailored letter
• An AI bedtime story generator — type in a character name and theme, get a story

Every one of these has the same architecture problems as the trip planner. Every one of them will benefit from the same lessons.