Leveraging AI · Week 2 · Video 1
Week 2 — Build

Reverse-engineer, and decide what to build.

Find proven ideas. Capture the designs you love. Lock in what you're building, then build it.

By the end of this, you'll know how to build the face of your web app, the kind of premium visuals people pay $10,000 for. Not basic, AI-looking landing pages. Premium, clean, and built around your expertise.

Today

We're doing four things.

Not sure of your idea yet? Do the validation step in Resources first, then come back.

Step 1

Find proven ideas + your competitors.

Don't guess. Look at what's already selling. There are marketplaces where founders list their products with real revenue attached, because to buy or sell a business, the seller has to show their numbers.

That makes them a window into what's actually making money, not just what sounds good.

How to use them: filter by revenue, study what sells and why, then reverse-engineer it and make it better. You're not copying. You're proving people already pay for this.

Open TrustMRR →
Also worth exploring (in your Resources): Acquire · Flippa · IndieHackers
Step 2

Capture the designs you love.

Capture the designs so Claude can build from them. The tool is SkillUI, free, no API key. It turns any site into markdown design files Claude can read. The trick: capture a few sites, then pull the best parts of each. You're not cloning anyone, you're assembling your own look.

Move 1 — Capture 2 or 3 sites you like. Paste into Claude:

Paste into Claude
Install the SkillUI CLI (npm package skillui, from github.com/amaancoderx/npxskillui) and run it on these sites, keeping each one's design files separate: [site one], [site two], [site three].

Move 2 — Compose. Tell Claude what to pull from where. Paste into Claude:

Paste into Claude
Using those SkillUI design files, build my landing page. Take the overall layout from [site one], the colors and fonts from [site two], and the pricing section style from [site three]. Use my content. Make reasonable choices and don't ask me questions.
Open SkillUI →
Step 3

Pick your platform type.

It does not have to be a video course. Pick what fits how your people actually learn or get results.

Course

Teach it step by step.

Quiz

Test and score them.

Practice

Reps and repetition.

Tool

Like a document analyzer.

Step 4

Nail your unique proposition.

This is your angle. Why you, and why this. It ties back to the positioning you nailed in validation.

What do you do, for whom, that nobody else in your space does quite the same way?

Lock that in. It shapes everything you build next.

Quick note

What about API keys?

Most of you won't need one to start.

Step 5

Write your spec, then build it out.

You've just decided everything: your competitors, your design, your platform type, and your angle. Now we hand all of it to Claude at once.

Your design files describe how it looks. Your spec describes what it does. Claude needs both.

Ask for the spec first, not the build. You want to read it and approve it before a single page gets made. That one habit is the difference between a platform and a mess.

Paste into Claude
Using the design files you just captured, write a one-page spec for my platform before you build anything. I'm building a [platform type] for [your audience]. My angle is [your angle]. I already have a login built with Supabase. List the pages, what each one does, and the tables you'd need. Show me the spec so I can approve it, then build it out.

Read what it gives you. If a page is missing or something's wrong, say so and have it fix the spec. Then let it build.

By the end of this video

Here's where you should be.

Next video: what great platforms do to keep people coming back. Now that you have something real on screen, you can decide which of those your platform actually needs.