That was the moment I decided to build something. I have no engineering background, so "build an iOS app" seemed overly ambitious. Today CouchRot is live on the App Store. AI does NOT make you a developer. It makes a non-technical person dangerously more capable, but there's an enormous gap between "the AI wrote me some code" and "I have a product in production that users actually trust." That gap is hours of debugging errors I didn't understand. Late nights learning what row-level security meant after I realized users could briefly see each other's data. More than one 2am moment. Stack: React Native + Expo, TypeScript, Supabase (Postgres, Auth, RLS across ~30 tables, 11 edge functions), OpenRouter for AI model routing, TMDB for movie data, RevenueCat for subscriptions, Sentry for errors, offline sync, push notifications. Built with a lot of Cursor and Claude/Cowork help. What it does: you tell it what you're in the mood for, rate some titles 0-100 to build a taste profile, and it gives you 5 picks tuned to what you actually like.
Non-Engineer Ships Full iOS App (CouchRot) to the App Store Using Claude/Cowork
A non-technical user built and shipped CouchRot (movie recommendation app) to the App Store using Claude/Cowork. Full stack: React Native + Expo, TypeScript, Supabase with RLS across 30 tables, 11 edge functions, OpenRouter, RevenueCat, Sentry, offline sync
Non-engineer who built and shipped a full iOS app using Claude/Cowork
Scenario
A non-technical user was frustrated by spending too long scrolling Netflix/Hulu/Max every night looking for something to watch. Despite having no engineering background, they decided to build an iOS app to solve the problem, using Claude/Cowork to bridge the technical gap.
Prompt
Help me build a movie recommendation iOS app from scratch. I have no coding experience.
Requirements:
- Users describe their mood ("date night", "need a laugh", "something mind-bending")
- Rate movies 0-100 to build a taste profile
- Get 5 personalized picks based on their actual preferences
- Stack: React Native + Expo, TypeScript, Supabase for backend
- Include subscriptions, error tracking, offline sync, push notifications
- Design philosophy: get users OFF the app fast and watching something
Expected Result
A complete iOS app built from zero and shipped to the App Store: **Tech stack:** - React Native + Expo + TypeScript - Supabase (Postgres + Auth + RLS across ~30 tables + 11 edge functions) - OpenRouter (AI model routing) + TMDB (movie data) - RevenueCat (subscriptions) + Sentry (error tracking) - Offline sync + push notifications **Real experience:** - 100+ hours of development total - Discovered a user data leak (RLS issue) and had to learn row-level security overnight - Multiple 2am debugging sessions - "AI does NOT make you a developer — it makes a non-technical person dangerously more capable" **Outcome:** CouchRot is live on the App Store — a complete journey from "AI wrote me some code" to "a product in production that users actually trust."