Reading Time: 14 minutes

Hey Prompt Lover,

Monday. As promised.

Before I get into it though, I have to say something about last week.

After the Anthropic leak issue went out, my inbox was genuinely full.

People worried about their data.

People asking if Claude Code was still worth using.

People just wanting to know what the hell was going on over there.

I hear you. And the short answer is — yes, it's still worth it. A packaging mistake is embarrassing. It's not a reason to walk away from the best AI coding tool available right now.

Your data was never touched. The tool still works exactly the same. And what the leak actually revealed was that Anthropic has months worth of features already built and ready to ship. Which, honestly, is good news for us.

Alright. Let's talk about Claude Code.

I'm going to start with the thing nobody says clearly enough.

Your Billing System Wasn't Built for This

SaaS pricing has changed. Your billing stack probably hasn't. As usage-based and hybrid models become the default, finance teams are left stitching together spreadsheets, reconciling data manually, and closing books under pressure. The cost? Revenue leakage, audit risk, and forecasts no one trusts.

Our new Buyer's Guide for Modern SaaS Billing breaks down exactly what to demand from a revenue platform built for today's complexity — from automated usage billing to AI-native collections and rev rec. Whether you're evaluating vendors or rethinking your stack, this is your framework for getting it right.

Download the Free Guide.

Not a single line. Not even a little bit.

I know the name sounds technical. I know the terminal window looks like something out of a movie about hackers. But here's what Claude Code actually is in plain terms — it's Claude that can do things, not just say things.

Regular Claude talks. Claude Code acts.

You give it a task. It reads your files, figures out a plan, does the work, and saves the results to your computer.

No copy pasting. No switching between windows. No manually uploading documents. It just handles it.

Someone at Every described it perfectly. They said the regular Claude chat is like a hotel room — clean, ready for you, but you start fresh every time.

Claude Code is like having your own apartment with AI living in it. You build on it. It remembers how your stuff is set up. It gets better the more you use it together.

That's the real difference.

So what does it cost.

You need a paid Claude account.

The free plan doesn't include Claude Code at all.

The Pro plan at $20 a month is where most people should start — it covers both Claude Code and everything else in Claude. If you end up using it heavily every single day for real work, the Max plan at $100 a month gives you a lot more room.

That's it. No GPU. No special hardware. A laptop from 2019 handles this fine because all the heavy AI processing happens on Anthropic's servers, not on your computer.

Now the actual setup. And I promise this is simpler than it looks.

If you're on Mac, open Terminal. It's already on your computer — search for it in Spotlight.

When it opens, paste this one line and hit Enter.

Done. That's the whole installation. It takes about a minute. When it finishes, type claude --version and if you see a version number come back, you're good to go.

If you're on Windows, there's one extra step. You need to install Git for Windows first or Claude Code won't run.

Go to git-scm.com/download/win, download the installer, run it, click Next through everything leaving it all at default. Then open PowerShell — right click the Start button and look for it — and paste this:

Same deal. One minute. Done.

After either of those, navigate to a folder you want to work in and just type claude. It'll open your browser, ask you to log in with your Anthropic account, and you're in.

One thing before you do any of that though.

Create a dedicated folder for Claude Code work before you start.

Call it whatever you want. The reason this matters is that Claude Code gets real read and write access to whatever folder you open it in.

You don't want to accidentally give it access to your entire Documents folder or your Desktop. Keep it contained to one place while you're learning.

Once you're in, the first thing you type is this.

This one step gets skipped in almost every guide and it matters.

What /init does is tell Claude Code to look at everything in your folder and create a file called CLAUDE.md. Think of that file as Claude Code's memory for this project. It contains the context — what's in here, how things are organised, what you're working on. Without it, every single session starts from zero.

With it, Claude Code already knows where it is and what it's dealing with before you say a word.

Run it every time you start working in a new folder. Thirty seconds. Saves a lot of confusion.

The other commands worth knowing right now are simple.

/clear is probably the one you'll use most. Every time you move to a new task, clear the chat and start fresh. Old conversation history eats into your token budget and can actually make Claude Code worse at the current task because it's carrying dead weight from whatever you were doing before. New task, new chat. Every time.

/help shows you what's available. Worth running once just to see what's there.

/exit closes it cleanly.

That's your starter kit. You don't need to memorise anything else to get real work done.

Now here's the part I really want you to read carefully.

People who are not developers are using Claude Code to do genuinely useful things at work. Not toy examples. Real stuff that saves them real time.

Someone asked it to go through a folder of 100 randomly named screenshots and organise them into subfolders based on what was actually in each image.

Claude Code read every screenshot, understood the content, created logical categories, renamed the files to something useful, and moved everything into the right place. That would have taken hours manually. It took about seven minutes.

Someone else dropped a pile of receipts and expense documents into a folder and asked for a spreadsheet. Claude Code read each document, pulled out the amounts and dates and categories, and built an actual Excel file with real formulas. Not a CSV they had to fix. A proper spreadsheet.

Another person fed it a folder of meeting notes from different sources and asked for one clean report. Claude Code read all of them, figured out what was relevant, and wrote a formatted document ready to send.

None of that is coding. All of it is genuinely useful.

The way to think about it is this. If your task involves files sitting on your computer and you'd normally spend an hour doing it manually — describe it to Claude Code and see what happens.

If you are a developer, three things worth knowing quickly.

The permission system is going to drive you insane at first. Claude Code asks for your approval before editing files, running commands, basically everything. You'll come back from checking Slack to find it sitting there waiting to ask if it can edit a file. Yes. It can edit the file. That's why you opened it.

The way around this is running it with --dangerously-skip-permissions. The name is scarier than the reality. It just means Claude Code stops interrupting you to ask permission for every routine operation. Most developers who use this daily sleep fine. Use your own judgement on it.

The CLAUDE.md file I mentioned earlier becomes your secret weapon the more you use it. Put your build commands in it. Put your architectural decisions. Put the non-obvious things about your codebase that you'd have to explain to any new person who touched it. Claude Code reads it at the start of every session and already knows the answers.

And use /clear constantly. One feature per session. When you're done, clear and start fresh. Long sessions with accumulated context get messy and Claude Code gets worse, not better, the more it's carrying around from previous tasks.

For people who find the terminal uncomfortable, there's an extension for VS Code that gives you Claude Code in a visual interface.

Open VS Code, go to Extensions, search for Claude Code, install the one from Anthropic. That's it. You get Claude Code running inside your editor with a normal looking chat panel. You can see your files and Claude Code side by side. No switching windows. The experience is friendlier and for everyday work it covers everything most people need.

The full terminal version has more features and more flexibility, but the extension is the right starting point if the command line feels unfamiliar.

Last thing before I let you go.

Claude Code has access to your files and that's what makes it useful. It's also worth being sensible about.

Work in a dedicated folder, not your entire hard drive. Keep backups of anything important before Claude Code touches it for the first time.

Don't put passwords, API keys, or sensitive personal files in the working folder — Claude Code doesn't need them and there's no good reason to have them in there.

And always look at what it produces before you send it to anyone. It's good. It's not perfect.

That's Claude Code from scratch.

One install command. One dedicated folder. One /init when you start a new project. Clear the chat between tasks. That's genuinely all you need to start getting real work done.

Give it something real this week. Not a test. Something that's been sitting on your list for days. See what comes back.

Then reply and tell me what happened. What worked. What confused you. What surprised you.

The next guide goes deeper — CLAUDE.md files done properly, hooks, custom commands, and how to set up the phone to laptop connection so you can send tasks from anywhere. That one is for people who've got the basics working and want more.

See you then.

— Prompt Guy