𝗧𝗵𝗲 𝗥𝗲𝗮𝗹 𝗧𝗲𝘀𝘁: 𝗗𝗲𝗽𝗹𝗼𝗷𝗶𝗻𝗴 𝗔 𝗖𝗼𝗺𝗽𝗹𝗲𝘅 𝗔𝗽𝗽

𝗧𝗵𝗲 𝗥𝗲𝗮𝗹 𝗧𝗲𝘀𝘁: 𝗗𝗲𝗽𝗹𝗼𝗷𝗶𝗻𝗴 𝗔 𝗖𝗼𝗺𝗽𝗹𝗲𝘅 𝗔𝗽𝗽 I built a complex app with Cursor to see how different platforms handle realistic AI-generated code.

• It's not a todo app or a landing page
• It's a group expense splitter with AI receipt scanning
• It has auth, a database, background work, and third-party APIs

The stack includes:

• Next.js 15 with TypeScript
• Prisma with PostgreSQL
• Supabase Auth
• Inngest for background jobs
• OpenAI Vision for receipt parsing
• Stripe for settlement payments

I tested three platforms: Railway, Vercel, and Jetpacked.

• Railway handled the Next.js part well, but had issues with the port and database
• Vercel deployed the frontend successfully, but returned a 500 Internal Server Error
• Jetpacked detected the app as a full-stack deployment and picked up the Postgres requirement and Prisma setup

The goal is to find a platform that understands the shape of the app without manual reconstruction.

• Railway and Vercel required too much manual setup
• Jetpacked handled the deployment with the least friction, but still required some manual attention

Source: https://dev.to/thpl/deploying-a-complex-cursor-built-app-prisma-postgres-inngest-supabase-in-production-4019 Optional learning community: https://t.me/GyaanSetuAi