From Prompt to Shipped Product in One Weekend

You don't need three months. You need a weekend and a credit card.

This is the actual sequence I've watched founders run dozens of times. It's not theory. It's hour-by-hour. By Sunday at 9 PM, you'll have a deployed app with Stripe wired, a friend's $5 test charge in your dashboard, and a Show HN draft in your clipboard.

Block the weekend. Tell your people you're offline. Read this once, then go.

The Weekend Cheat Sheet

Friday, 7 PM: Validate (90 minutes, not weeks)

Open a notebook. Write the app idea as one sentence:

"A tool that does X for people who are Y, charged at $Z/month."

If you can't fit it in that sentence, the idea isn't ready yet. Reshape it until it fits.

Then do 60 minutes of validation. Not user interviews. Not surveys. Just three things:

1. ·Search Reddit for "anyone know a tool that does X." If 10+ threads exist, real demand.
2. ·Check if 2-3 ugly competitors exist. Competitors are good. They prove someone pays.
3. ·Text 3 friends who fit profile Y the one-sentence pitch. If two say "I'd try that," go.

Don't do user research. You're not building a startup. You're building a $30K-MRR-someday SaaS that you'll ship in 48 hours. Validation past this point is procrastination.

Friday, 8:30 PM: Sketch Three Screens

On paper or in Figma, sketch only three screens:

1. ·Landing page. Headline, one paragraph, signup button. That's it.
2. ·The main app screen. The one place users will spend 90% of their time.
3. ·The settings / billing page. Where they manage their subscription.

Don't sketch the dashboard. Don't sketch the admin panel. Don't sketch the onboarding modal. Those are tomorrow's problems. Three screens. 30 minutes.

Friday, 10 PM: First Cursor Prompt

Open Cursor. Open a new folder. Open chat.

First prompt, copy-paste-able:

"Create a Next.js 14 app with TypeScript, Tailwind, and shadcn/ui set up. Add three pages: landing (/), app (/app), and settings (/settings). Use App Router. Add a basic header with the app name and a sign-in button. Make the landing page look like a clean Vercel-style hero with one headline and a CTA. Use Switzer or Inter as the font."

Wait. Let Cursor scaffold it. When it's done:

npm install
npm run dev

If localhost:3000 shows your landing page, you're good. If it errors, paste the error back into Cursor and let it fix itself. This loop is the entire skill of vibe coding. Trust it.

Bedtime: midnight. Don't push through. Tomorrow is long.

Saturday, 9 AM: Build the Core Loop

Coffee first. Then the prompt that matters most all weekend:

"On the /app page, build the main functionality: [describe it in plain English in 3-5 sentences]. For now, store everything in component state. We'll add the database in a few hours. Make it work end-to-end with fake data so I can see the full user flow."

The trick: build with fake data first. Don't wire the database yet. Get the screens looking right, the buttons doing the right things, the empty states polished. You can swap fake data for real data in 30 minutes once it works.

Spend until about 1 PM on this. Lunch break, 30 minutes, away from the screen.

Saturday, 2 PM: Wire Supabase

Sign up for Supabase. Create a new project. Free tier. Takes 2 minutes.

Cursor prompt:

"Add Supabase to this project. Set up auth with email/password and Google OAuth. Create a users table and a [your-main-entity] table with row-level security so users only see their own data. Replace the fake data in /app with Supabase queries. Add a sign-in / sign-up flow that redirects to /app on success."

This will take 60-90 minutes including the debugging. The first time you watch a real signup flow work end-to-end on your own app is the moment the weekend becomes real.

By 4 PM, you should be able to sign up as a fake user, see an empty app, create something, and see it persist on refresh. If yes, you're on schedule.

Saturday, 6 PM: Stripe Checkout

Sign up for Stripe. In test mode, create one product with one price (start at $9/month, you can change it later).

Cursor prompt:

"Add Stripe Checkout for subscription. When the user clicks 'Upgrade' in /settings, create a Checkout session and redirect them. Handle the success and cancel URLs. Add a webhook handler at /api/webhooks/stripe that updates a subscription_status column in the users table when subscriptions activate, update, or cancel. Use the Stripe customer portal for managing subscriptions."

Stripe's docs are good and Cursor knows them well. This is a 60-90 minute task in test mode.

Saturday, 9 PM: Deploy and Test With a Real Person

Push to GitHub. Connect the repo to Vercel. Click deploy. Done.

Add your environment variables in Vercel (Supabase URL, anon key, Stripe keys, the webhook secret). Redeploy.

Visit the live URL. Sign up with your personal email. Click Upgrade. Pay with the Stripe test card 4242 4242 4242 4242. Confirm your subscription status updated in Supabase.

Then text one friend the link. Don't write a paragraph explaining it. Just: "I built this in a weekend. Try it. Tell me what breaks."

When they hit a bug at 10 PM (they will), fix it in Cursor, push, Vercel redeploys in 90 seconds. This is the loop. This is shipping.

Bed by midnight again. Tomorrow is the cleanest, easiest day.

Sunday, 11 AM: Polish

Three things, no more:

1. ·Landing page copy. Rewrite the headline 5 times. Pick the best one. The headline matters more than anything else on the page.
2. ·The empty state on /app. First-time users see this. Make it obvious what to do next. Add one button, one sentence.
3. ·A 2-minute Loom on the landing page. Record yourself using the app. Embed it under the headline.

That's it. Don't add features. Don't redesign. Polish only what 95% of users will see in their first 30 seconds.

Sunday, 4 PM: Switch Stripe to Live Mode

Get your live Stripe keys. Update Vercel env vars. Test with a real $1 charge to your own card (you can refund it). Confirm it works end-to-end with real money.

You now have a real, charging, deployed SaaS app. Less than 48 hours since Friday night.

Sunday, 6 PM: Ship the Post

Pick one launch surface:

• ·Show HN if it's a developer tool
• ·Reddit (the relevant subreddit) if it's a consumer or niche B2B tool
• ·X / Twitter if you have any audience there
• ·Indie Hackers if you want feedback from other builders

Title format: "I built [thing] in a weekend so I could [outcome]." Honest, plain, no exclamation marks. Link to the app. Reply to every comment for 4 hours.

You won't get 10,000 visitors. You'll get 80-300. Of those, 3-10 will sign up. Of those, 1-3 will pay. That's $9 to $27 of MRR by Sunday at midnight. That's a real business. From there, it's just a matter of doing this loop again next weekend with what you learned.

Sunday, 9 PM: close the laptop. You shipped.

Now schedule next weekend.