Build a SaaS and get a stranger to pay for it in 4 days.

You're about to run cold traffic at this page, so don't invent a layout. Go to mobbin.com, find a landing page you like from a top app, screenshot it, and have Cowork reverse-engineer why it works. I spent most of day 1 iterating here, on purpose. Nothing else matters if people never get past this page.

Here's a screenshot of a landing page I like: [ATTACH SCREENSHOT]. Analyze its layout and sales psychology: section order, headline formula, social proof placement, CTA strategy, objection handling.

Then turn that analysis into a PRD that applies the same structure and psychology to my app [NAME], using my brand kit. I'll be sending cold paid traffic to this page, so optimize for first-visit conversion.

Monad's promise: find the business you're built for, based on your chart. On signup, a low-friction form collects birth info, then the LLM generates a free "founder archetype" reading. The free reading has one job: be scary accurate. It builds trust, overcomes skepticism, hits their ego, and gives a dopamine hit. THAT is the moment you show the paywall, not before.

You generate hyper-personalized [READING TYPE] for [AUDIENCE]. Input: name, birth date, time, and location. Compute their [zodiac sign, life path number, Chinese zodiac].

Output a reading that:
1) opens with a specific, flattering, uncannily accurate observation about how they operate,
2) names their archetype with a strong title,
3) connects every trait to a concrete [BUSINESS/OUTCOME] recommendation,
4) never hedges or sounds like a horoscope column.

Confident, direct, premium tone.

Expect to spend a few hours iterating on this until outputs feel personal instead of generic. That iteration IS the product.

Create a free Whop account, go to whop.sh, copy the install link, and add the Whop CLI to Claude. Then one prompt creates the products and embeds checkout.

Using the Whop CLI, create two pricing options for [NAME]:
1) $8.99/week with a 3-day free trial
2) $98/year

Embed the Whop checkout modal on the site at the paywall step of onboarding.

Research the payment flows of the top 5 [CATEGORY] apps: paywall placement, pricing presentation, trial framing, urgency and conversion tactics. Turn the findings into a PRD for my app's payment modal that I can hand to Claude Code, using my brand kit.

The paid offer ($8.99/week) is daily, weekly, and monthly readings: a living product people check every morning, not a one-time report. Two monetization tactics did the heavy lifting:

The blur. Free users can read half of the daily reading. The rest is blurred with a trial CTA. The biggest apps in the world use this because it works: give a taste, dangle the carrot.

The upsell. An AI chat that answers questions using their chart data. 15 messages a day on the paid plan, then $2.99 for extra credits. This raises revenue per user above the subscription price.

With the hook, the plan, and the upsell in place: buy a domain on Vercel, tell Claude to deploy. End of day 2, a fully functional SaaS, live on the internet.

I chose paid ads over content because you get messaging data fast and it's the quickest path to customers. Whop generates the creatives AND runs the Meta campaign, no Meta account needed. Upload a logo and banner, enter your business name, and Whop spins up a Facebook page for you.

Look at my landing page at [URL]. Using Whop's ad generator, create 5 ad creatives that match its messaging and visual style.

[Review them, then:] Generate 15 more variations so I have a pool to test.

Set up the ad campaign in Whop on Meta using the creatives we generated. Budget: $100/day. Target ONLY the US, UK, Canada, and Australia.

Day 4 morning: first user entered their credit card. The raw funnel: 100+ visitors, 32 generated a chart, 1 converted. That's validation, not profit. $40 of spend for one $8.99/week customer is a $40 CAC when it needed to be under $20. Now the real work: find the leak.

Here's a screenshot of my funnel conversion rates at each step: [ATTACH]. Compare each step against industry-standard benchmarks for consumer subscription apps. Tell me which step is furthest below benchmark and the 3 most likely causes.

Mine showed onboarding completion 20% below benchmark, and 75% of traffic was mobile. I tested my own product on my phone and found 2 bugs in 5 minutes: the location field ignored manual entries, and the time field silently required a colon. Claude fixed both, completion rose instantly, 3 more customers came in that day, and CAC dropped to $25.

Most users won't start a trial in their first session, and you already collect their email in onboarding. So send each lead a custom daily reading every morning to pull them back. I set this up at 3am with a Resend account and the Zapier MCP. It adds token cost but the conversion lift is worth it.

Connect Resend via the Zapier MCP. Create an automated daily email campaign: every morning, generate a custom [READING/CONTENT] for each lead using their [DATA COLLECTED IN ONBOARDING] and email it to them with a CTA back to the app. Write the template to match my brand voice.

Day 5 hurt: $100 of spend, zero trials. The cause was an authorization bug. Nobody was hitting the paywall at all. Check your paywall actually fires, every single day, while money is being spent. I fixed it, made some UI tweaks, and waited.

Day 6: 7 new users in 24 hours on the final $100. That's a $15 CAC, under the $20 target Claude calculated from my pricing model. The challenge goal was validation, and the idea validated in under a week.

The 3 mistakes that cost me ~$120

1. No geo targeting on launch: $20 gone to junk traffic.
2. Never tested onboarding on mobile: days of suppressed conversion from 2 tiny bugs.
3. Silent paywall auth bug: $100 of spend with nobody seeing a price.