Vibe Coding for beginners: Claude Code, Lovable & Figma Make

Go From Zero to a Working, Published App. No Jargon. No Computer Science Degree.

Most AI courses hand you a tour. You watch someone click around, you feel inspired for a day, and you never build the thing.

This one ends with your app live on your own domain.

Not a mockup. Not a pretty screen that does nothing when you click it. A real application with a working backend, a database, a connected AI, and a URL you can send to another human being.

This is a tested system, not a theory

Everything in this course is a system I run on real products that are live right now, with real users and real paying customers.

That matters more than it sounds. It means every prompt in here has been run hundreds of times, not written once for a video. It means the workflows have already failed, in the specific ways they fail, which is how I know where they break and can warn you before they break for you. It means when I tell you to do something in a particular order, it's because I did it in the wrong order first and paid for it.

You are not getting someone's best guess. You are getting the exact system, in the exact sequence, that produced working software.

I will not bury you in terminology

Here is the thing nobody says out loud. Most technical teaching is bad on purpose. Vocabulary gets used as a gate. If you don't already know what a repository or a migration or an environment variable is, you're made to feel like you're behind before you've started, and you quietly close the tab.

I don't do that.

Every term in this course gets explained the moment it appears, in plain words, once. If a step needs a command typed into a black window, I show you exactly what to type and what should happen next. You never need to know why it works to make it work. You'll pick that up naturally, later, in the order that suits you rather than the order a syllabus decided.

You bring an idea. I handle the vocabulary.

What you actually walk out with

By the last lecture you will have built and published a functioning SaaS product. You'll have a research method you can point at any topic, a library of reusable AI skills you own forever, and a master prompt system that turns a vague idea into something buildable in an afternoon.

You'll also have my assets. This is the part most people enroll for.

• The Prompting Guide. A pre-engineered instruction set. Upload it and your AI stops behaving like a chatbot and starts behaving like a senior builder.

• The PM UX Manual. Modern UI patterns and design psychology, written as an algorithm your AI can follow. Feed it in and your screens stop looking like a template.

• The Lovable Bible, the UX Kit and the Figma Bible. The specific instructions that get each tool to produce work you can ship instead of work you have to redo.

• The Master Prompt. The six-part structure I use to go from idea to working app. You'll build your own version of it during the course, using your own idea.

What you'll learn to do

Research like you have a team behind you. Most people write prompts from whatever they half-remember. You'll learn to research a topic properly first, using Gemini Notebook to pull real sources with citations you can click back to. Then you turn that research into prompts that actually know something.

Build AI skills you write once and use forever. This is the highest-leverage thing in the course. You'll learn to package a method into a Claude skill, then build a whole library of them: market analysis, UX decisions, scope review. Write it once. Never write it again.

Turn a vague idea into a buildable spec. You'll run your skills together against one app idea and watch each one catch what the others missed. The output is a master prompt with six parts and full annexes. That document is what makes the build phase feel easy, because all the thinking is already done.

Ship it with Lovable, then take it further. Feed the master prompt in piece by piece and watch the app appear. Then connect it to the Claude API so your app can read and analyze documents your users upload.

Go past the prototype ceiling. This is where most AI courses stop and most people get stuck. You'll connect Lovable to GitHub and GitHub Desktop, install Claude Code, and work on the real codebase locally. Backend work. The database. A custom domain. Every step shown, every command given to you. This is the difference between a demo and a product.

Build local skills from public GitHub repos. Half the best material in the world is already sitting on GitHub for free. You'll learn to find the good stuff, set up your project so Claude Code stops forgetting what you're building, and turn real code into skills you can reuse.

What we build, section by section

1. The prompting foundation and your asset pack. Download the Prompting Guide, the Lovable Bible, the UX Kit and the Figma Bible, then configure your AI so it stops behaving like a search engine. This is the setup that makes everything after it faster.

2. Strategy and the Master Prompt. The product manager's mindset. How to sync what you write with what you show, so the AI stops guessing at your intent. Ends with the Master Prompt, the structure the whole course runs on.

3. Figma Make vs Lovable. An honest comparison, tested in practice rather than argued in theory. What each one is genuinely good at, and how to pick without wasting a weekend on the wrong one.

4. Your first real prototype. Build a spacecraft monitoring interface end to end in Figma Make. Map the user flows first, then connect them into something that behaves like a product instead of a pile of disconnected screens.

5. Your personal website, live. Brand, strategy, master prompt, then zero to one in Lovable. Plus the pitfalls that bite everyone the first time they push a prototype toward being a real product.

6. Research like you have a team. Gemini Notebook end to end. Research any topic with real cited sources, then use the same method to build the foundation of a market analysis skill.

7. Claude, projects and skills. What actually differs from ChatGPT. Projects, skills, and the skill that writes skills. You'll build a library, run it against your own app idea, and assemble a six-part master prompt with annexes.

8. Ship it with Lovable, GitHub and Claude Code. Build the app from your master prompt, wire in the Claude API, sync to GitHub, then go local with Claude Code for backend, database work and a custom domain. Finish by publishing.

9. Test your knowledge. A final quiz that pins down what you learned.

Who this is for

• Founders who need an MVP without spending months and a budget on an agency.

• Product managers who are done writing tickets and waiting for a sprint that never has room.

• Designers who want to stop resizing rectangles and start shipping the thing.

• Solopreneurs validating an idea this month instead of next quarter.

• Career switchers who need a portfolio that proves they can build, not talk about building.

• Anyone who has been told they need to "learn to code first" and suspects that isn't true any more.

Who this is not for

If you want a computer science education, this isn't it. I'm teaching you to build working software, not to become an engineer. Those are different goals and this course is honest about which one it serves.

If you want to watch videos without opening a single tool, you'll finish with exactly what you started with.

And if you need to understand every line of code before you'll trust it, you'll find this course frustrating. The whole method depends on trusting a tested system while you're still learning why it works.

You need an idea and about an afternoon a week. That's the whole entry fee.

Why I made this

The gap between people who can build and people who can only describe what they want has never been wider, and it has never been easier to cross.

I crossed it. I run a digital workplace organisation by day and I've built and launched real products on the side using exactly what's in this course. Not in theory. On weekends, around a family, with the same 24 hours you have.

The tools got good enough a while ago. What's missing for most people isn't talent or time. It's the words to say to them, in the right order.

I'm going to give you the words.

Enroll now and build the thing you keep describing to people.