From Zero to Working Prototype in 48 Hours with Replit

How I built a booking app for tattooers over the weekend, not because I’m a developer, but because I don't have to be.

I started out as a programmer and I used to sling C like a badge of honor. Then life happened—startups, management, a divorce, two kids—and the IDE gathered dust. My “coding” for the past decade was the odd landing page, a CSS tweak, nothing more. So when I spun up a full booking-and-payment app for tattoo artists in a single weekend, I felt like I’d stepped through a portal. The portal’s name: Replit.

Then a couple of months ago I read a story. Reid Hoffman asked Replit to rebuild LinkedIn with a single prompt. And it actually worked.

I opened Replit out of curiosity, played around for a couple of hours, then closed the tab. Three days ago I came back. Not to experiment, but to build.

Tattoo artists spend way too much time dealing with DMs, no-shows, and flaky clients. What they need is simple. One link where clients can view availability, book an appointment, and get a confirmation instantly.

So I opened a blank repl. Described the product. And let the machine take it from there.

48 Hours. A Real App. Working End to End

I opened a blank repl on Thursday night, typed a paragraph—“calendar like Calendly, Google OAuth, Stripe deposits, Postmark notifications”—and watched the Agent sketch my backend, wire Stripe, set up calendar integration, and scaffold a React front-end before I finished my coffee.

Saturday I argued with it about time-zone math, cursed at it when it broke things, and pushed for UI polish. And I barely touched a compiler.

That two day sprint felt less like coding and more like product conversation. And it’s a small taste of the tectonic shift now rolling through software.

Clients could book appointments through a clean mobile-friendly calendar. Artists had dashboards to view bookings, track stats, manage clients, and customize their availability with a complete onboarding flow.

Artists could sign up, reset passwords, upload profile photos, and edit settings. The entire booking system handled time conflicts, tracked services, and sent confirmation emails to clients using Postmark. Fast, reliable, and ready.

Stripe is connected and configured. Part of our roadmap is to let artists accept payments and manage deposits directly through the app. The infrastructure is already there and solid.

And I didn’t write code the old way. I wrote prompts. Debugged in plain English. Fixed what broke. Argued with the AI when it ignored instructions. Iterated. Shipped.

Replit Is Changing the Rules

This wasn’t a weekend toy. It’s not a mockup with fake data. It’s a real product built with actual logic and working integrations, even if it hasn’t been launched yet.

Replit didn’t just speed things up. It killed the prep work. No local installs, no AWS console, no backend wrangling. Just an AI agent and me iterating in the same tab.

Is it perfect? Not even close. Sometimes it rewrites working code for no reason. Sometimes it breaks things and confidently tells you it fixed them. But even when it’s wrong it’s fast. And when it works it feels like magic.

This isn’t about productivity. This is about power.

Reid Hoffman Asked Replit to Clone LinkedIn and It Delivered

Replit didn’t just help me build. It helped Reid Hoffman rebuild LinkedIn. One prompt. Functional prototype. This isn’t fiction. This is now.

A quarter of the latest YC batch uses AI to write nearly all their code. Replit went from ten million to one hundred million ARR in six months. Zillow is letting non-engineers push updates to production. PMs are building tools. Tattoo artists will be booking clients with software written by someone who hadn’t touched code in years.

The gatekeepers are gone. The hard part isn’t writing code anymore. It’s knowing what to build and having the guts to start.

This Is the New Game

This wasn’t a prototype to show off in a deck. It’s a prototype product built to solve a real problem. Every core flow is functional. Every feature is intentional. I didn’t cut corners or fake progress. I built something usable, end to end, in a weekend—not to impress anyone, but to prove what’s now possible when you remove the friction.

And I built it. No dev team. No designer. No permission.

If you’re still sitting on an idea waiting for someone else to build it for you, forget it. That era is done.

You don’t need a co-founder. You don’t need funding. You don’t need any of that anymore. You can talk? Good. Talk to the agent. Debate with it. Curse it out when it screws up. Break things. Fix them. And ship.

This is how software gets made now. Wait around, and you’re already behind.