It's Just a Text Window

A terminal is a text-based window where you type commands to your computer. That's it. No magic, no hacking, no Matrix-style green characters raining down the screen. It's literally a window where you type words and your computer responds.

Think of it this way: the terminal is like texting your computer. You send a message (a command), and your computer responds with a result. You're already used to this concept from WhatsApp or Messenger — you type something, you get something back. Same idea.

The only difference is that instead of texting a person, you're texting your computer. And your computer is very literal — it does exactly what you ask, nothing more, nothing less. It won't guess what you mean. It won't assume. It just does exactly what you type.

Why Not Just Use the Regular Interface?

Great question. You're probably wondering: "I've been using my computer with clicks and buttons for years. Why would I ever type commands?" The answer is simple: some powerful tools only work through the terminal. Claude Code is one of them.

Think of it like driving. You usually drive automatic — easy, comfortable, gets you where you need to go. But some race cars only come in manual. The terminal is like a manual transmission — it gives you access to tools and capabilities that don't have a "click" version. Claude Code is one of those tools. It's built for the terminal because the terminal gives it direct access to your files, your APIs, and your system.

The good news? You don't need to become a terminal expert. You'll type a few commands during setup, and after that, Claude Code handles everything through natural conversation. You talk to it in plain English inside the terminal window.

What It's Called on Your Computer

• Mac: It's called "Terminal" — a built-in app that comes with every Mac.
• Windows: It's called "Command Prompt" (cmd) or "PowerShell" — also built-in. There's also "Windows Terminal" which you can download from the Microsoft Store for a nicer experience.
• Linux: Usually just called "Terminal" — if you're on Linux, you probably already know this.

How to Open It

What You'll See

When you open the terminal, you'll see a window — dark or light background — with a blinking cursor waiting for you to type. That's it. No buttons, no menus, no icons. Just a cursor.

It might show something like this before the cursor:

mazin@MacBook-Pro ~ %

Or on Windows:

C:\Users\Mazin>

This is just your computer telling you who you are and where you are (which folder you're in). You can completely ignore it. The important thing is the blinking cursor — that's where you type.

What If Something Goes Wrong?

The terminal can't break your computer. If you type something wrong, you'll get an error message — that's it. No harm done. Just read the error and try again. The most common "mistake" beginners make is typos. If you type LS instead of ls, you'll get a "command not found" error. That's normal. Commands are case-sensitive on Mac/Linux (lowercase matters). Just retype correctly and move on.

If you ever feel "stuck" in the terminal — like it's not responding or something weird is happening — just press Ctrl + C. That's the universal "cancel" button in the terminal. It stops whatever is running and gives you your cursor back.

Common Commands (Don't Memorize — Just Know They Exist)

You don't need to memorize any of these. You'll copy-paste the few commands you need during setup. But here are the most basic ones so they don't surprise you:

• ls (Mac/Linux) or dir (Windows) — lists the files in your current folder. Like opening a folder and looking at what's inside.
• cd folder-name — goes into a folder. Like double-clicking a folder to open it.
• cd .. — goes back one folder. Like clicking the "back" button.
• mkdir name — creates a new folder. Like right-clicking and choosing "New Folder".

That's the entire list of commands you need to know. Four commands. Everything else, Claude Code handles for you through natural conversation.

claude.ai — The Website (Chat)

You already know this one. You go to claude.ai in your browser, you type a question, and Claude responds with text. It's a conversation. It's helpful. But it has hard limits:

• It can only respond with text — words on a screen
• It has no access to your computer, your files, or your folders
• It has no access to Meta Ads, Shopify, or any external platform
• It can't execute code, make API calls, or browse the web
• It's like texting a really smart friend who gives you advice but can't actually DO anything for you

Claude Code — The Terminal App

Claude Code is a completely different product. It runs in your terminal (the text window we just learned about), and it lives on your computer. This gives it superpowers:

• It can read and write files on your computer
• It can make API calls to Meta, Shopify, and any other platform
• It can browse the web (using Playwright — more on this later)
• It can execute code — run scripts, process data, generate reports
• It can interact with external services — create campaigns, pull analytics, modify products
• It's like having that smart friend physically in your office, sitting at your computer, with access to everything — and asking your permission before every action

Why This Matters for Media Buying

Here's where it gets real. Let's compare what happens when you ask for the same thing:

• Chat can TELL you how to create a campaign — you still have to go to Ads Manager and do it manually, step by step
• Code can actually CREATE the campaign via Meta's API — it's done, the campaign exists
• Chat can DESCRIBE your performance trends — you still need to open Ads Manager, export data, and analyze it yourself
• Code can PULL your real data via the API, analyze it, and present a formatted report — ready to read
• Chat can SUGGEST which ad sets to scale — you still decide and manually adjust budgets
• Code can CHECK performance, IDENTIFY winners, and INCREASE budgets via the API — with your approval

See the pattern? Chat gives you information. Code takes action.

A Day-in-the-Life Comparison

Let's make this even more concrete with a real scenario:

Morning check with Claude Chat: "How are my campaigns doing?" Claude responds: "I don't have access to your ad account, but here's how to check: open Ads Manager, go to Campaigns tab, look at your ROAS column..." You spend 15 minutes navigating Ads Manager, exporting data, and making sense of it.

Morning check with Claude Code: "How are my campaigns doing?" Claude Code runs your Performance Monitor Skill, pulls your data via Meta's API, analyzes all active campaigns, and presents a report: "3 campaigns running. Top performer: Summer Sale at 4.2x ROAS. Underperformer: Brand Awareness at 0.8x ROAS — recommend pausing. 2 ad sets ready for scaling." Done in 30 seconds.

That's the difference. Not just in capability, but in time. Every manual task you do today becomes a conversation with Claude Code tomorrow.

The Plans Explained

Claude Code requires a paid subscription to Anthropic. There are two tiers that matter:

• Claude Pro ($20/month): Gives you access to Claude Code with moderate usage limits. This is perfect for learning, experimenting, and running agents a few times per day. For most people starting out, this is more than enough.
• Claude Max ($100/month): Same access to Claude Code, but with significantly higher rate limits. This is for when you're running multiple agents daily across real client accounts and need heavy, uninterrupted usage.

Which Should You Pick?

Start with Pro. Learn the system, build your first Skills, run your first agents. When you find yourself hitting rate limits because you're running agents daily for multiple accounts, upgrade to Max. Don't overspend before you know the system.

The Real Comparison

Think about what the alternative costs: hiring a junior media buyer runs $500 to $2,000 per month. An AI assistant that works 24/7 costs $20 to $100 per month. It doesn't take sick days, doesn't forget things, and doesn't need training on how to use Ads Manager — it already knows.

The Restaurant Analogy

Imagine you're sitting at a restaurant. Here's how it works:

• YOU = your AI agent (the one who wants something done)
• THE MENU = API documentation (lists everything you can order, with descriptions and rules — "no substitutions on the grilled chicken")
• THE WAITER = the API itself (takes your order to the kitchen, brings back the result)
• THE KITCHEN = the platform's system (Meta, Shopify, etc. — does the real work behind the scenes)

You never walk into the kitchen. You never touch the stove. You read the menu, tell the waiter what you want, and you get your food. That's exactly how APIs work.

Translating the Analogy to Tech

Here's the same analogy in technical terms — don't worry, they map perfectly:

• "Ordering food" = making an API request (sending specific data to the platform)
• "Getting your plate" = receiving an API response (the platform sends data back)
• "The menu items" = API endpoints (specific things you can ask for — "create campaign", "get ad performance", "update product price")

The Two Main Actions

Just like at a restaurant, there are really only two things you do:

• READ (GET) — Looking at the menu. You're asking for information without changing anything. "Show me my campaign performance." "What products are in stock?" "How many orders came in today?"
• ORDER (POST) — Placing an order. You're asking the system to create or change something. "Create a new campaign." "Update this product's price." "Pause this ad set."

That's it. Every API interaction in the world is either "show me something" (GET) or "do something" (POST). Reading the menu or placing an order. When your AI agent communicates with Meta or Shopify, it's doing one of these two things every single time.

Documentation = The Rulebook

Every platform that offers an API also provides documentation (often called "docs"). Documentation is a detailed guide that explains:

• What you CAN request — which endpoints (menu items) are available
• What format to use — how to structure your request (like specifying "medium rare" for your steak)
• What responses to expect — what the platform will send back
• What errors might happen — "we're out of the salmon" (the API equivalent of an error message)
• What permissions you need — some menu items are only for VIP members (restricted API endpoints)

Where to Find Docs

Every major platform has a developers section with their documentation:

• Meta (Facebook/Instagram Ads): developers.facebook.com
• Shopify: shopify.dev
• Google Ads: developers.google.com/google-ads
• TikTok Ads: business-api.tiktok.com

Do You Need to Read Docs Yourself?

No. And this is a critical point. You don't need to read documentation yourself. Claude Code and the Skills files (covered in Part 5) handle this for you. The Skills already contain the exact API endpoints, parameters, and formats that Claude needs.

But knowing that docs exist helps you understand a powerful feature: the Self-Learning Protocol. Every Skill in this framework includes instructions that tell Claude: "If something doesn't work, go read the latest official documentation and figure out the new way." This means your Skills automatically adapt when platforms update their APIs.

What a Docs Entry Looks Like (Simplified)

Just so you recognize it if you ever see one, here's a simplified version of what an API documentation entry looks like:

Endpoint: POST /v21.0/act_{ad_account_id}/campaigns

Description: Creates a new ad campaign.

Parameters:
  name          (required)  — Campaign name
  objective     (required)  — OUTCOME_SALES, OUTCOME_TRAFFIC, etc.
  status        (required)  — PAUSED or ACTIVE
  special_ad_categories (required) — [] or ["HOUSING", "CREDIT"]

Response:
  { "id": "123456789" }

Errors:
  100 — Invalid parameter
  190 — Access token expired

This is the "menu item" for creating a campaign. Your Claude agent reads this kind of information and knows exactly how to "order" from Meta's API. You never need to deal with this directly.

The VIP Wristband Analogy

Imagine you're going to an exclusive club. At the door, they give you a wristband. That wristband proves you have permission to be there — without it, security won't let you in.

A Meta Access Token works the same way. It's a long string of characters (like a password) that your AI agent shows to Meta every time it wants to do something. Without it, Meta says: "Who are you? Access denied." With it, Meta says: "Welcome, you have permission to manage this ad account."

Types of Tokens

Meta has three types of access tokens, each with different "expiration dates" on the wristband:

• Short-lived token (1 hour): Expires in 60 minutes. Useful for quick testing — "let me try one API call to see if this works." Not useful for daily operations.
• Long-lived token (60 days): Lasts about 2 months. This is what most people use day-to-day. You'll need to renew it every 60 days.
• System User Token (never expires): For production systems that run 24/7. More complex to set up but never needs renewal. This is the goal for serious operations.

How to Get Your Token

Here's the step-by-step process. It looks long, but each step takes 30 seconds:

Security: Keep Your Token Secret

Your access token gives full control over your ad account. Anyone with your token can create, modify, or delete campaigns and spend your money. Here's how to keep it safe:

• Never share your token in chat messages, emails, or screenshots
• Never commit it to a public code repository (like GitHub)
• Store it only in a .env file on your local computer (we'll set this up in the setup section)
• If you think your token was exposed, immediately revoke it and generate a new one

Same Concept, Different Platform

Just like Meta has access tokens, Shopify has them too. The difference is how you get them: Shopify uses something called "Custom Apps" to grant API access to your store.

Think of it like giving a new employee specific access rights at your company. You don't give them the master key to everything — you decide exactly what they can see and do. "You can view inventory but you can't change pricing. You can read orders but you can't issue refunds." That's what Shopify's access scopes do.

Access Scopes = Permission Levels

When you create a Custom App, you choose exactly what it's allowed to do. Common scopes for media buying agents include:

• read_products — See your product catalog (names, prices, images)
• write_products — Update product details (prices, descriptions)
• read_orders — See order data (for revenue reporting)
• read_inventory — Check stock levels (critical for ad management — don't advertise out-of-stock products)
• read_analytics — Access store analytics and reports

How to Create a Shopify Custom App

Browser Automation, Explained

Playwright is a browser automation tool. In simple terms, it lets Claude Code open a web browser, navigate to websites, read pages, click buttons, and extract information — all without you touching anything.

Imagine having an invisible browser running in the background. Claude tells it: "Go to this URL. Read what's on the page. Bring the information back to me." That's Playwright.

Why This Matters for Your Skills

Here's the real power move. Every Skill in this framework includes a "Self-Learning Protocol". This protocol tells Claude: "If something doesn't work, don't just give up. Go read the official documentation using Playwright and figure out the new way."

This means your Skills never go stale. When Meta updates their API (which happens multiple times per year), your Skill doesn't break permanently. Instead:

1. Claude tries to execute an API call using the method described in the Skill
2. The call fails because Meta changed something
3. The Self-Learning Protocol activates
4. Claude uses Playwright to open Meta's official documentation website
5. Claude reads the updated docs and finds the new correct format
6. Claude retries the call with the updated method
7. It works

It's like having an employee who, when something doesn't work, automatically goes and reads the updated manual instead of just saying "I don't know, it's broken."

Other Uses for Playwright

Beyond self-learning, Playwright has other powerful uses for media buying:

• Reading competitor ad libraries: Claude can browse Facebook Ad Library and pull competitor creatives
• Checking product pages: Verify that landing page URLs are working before running ads
• Extracting social proof: Read product review counts or ratings from your store
• Monitoring price changes: Check competitor pricing on their websites