MVP for an AI startup in the field of advertising video generation

Frontend: Lovable | Backend: n8n | Supabase | Vercel | Stripe | Veo | NanoBanana | KIE
## Task

An American startup approached me with the task of developing an MVP AI platform for generating advertising video content.

The project involved:

A two-sided platform model with various types of users

Integration of several AI engines

Payment infrastructure

Management of digital entities and access rights

Asynchronous processing of media generation

Scalable SaaS architecture

It was necessary to design and implement the MVP from scratch — from architecture to production deployment.

# Architectural solution:

I designed the system as follows:

Frontend: Lovable (React-based UI)

Backend orchestration: n8n

Database & Auth: Supabase

Hosting: Vercel

Payments: Stripe

AI Engines: Veo, NanoBanana, KIE

Implemented logic

The platform includes a complex interaction model between several types of users and digital entities.

# The following were implemented:

- Role-based authentication and access rights management

- A system for managing digital presets and entities

- Asynchronous orchestration of AI processes via n8n

- Integration of multiple AI providers

- Real-time generation status processing

- Stripe (checkout + webhooks)

- Transaction management and access to functionality

- Supabase Storage for media

- Error handling and logging

Project challenges:

- Synchronization of frontend and backend states

- Asynchronous AI generation processes

- Separation of access logic between user types

- Scalable SaaS architecture

- Integration of multiple AI providers into a single system

# Result:

Working MVP

Production-ready architecture

Prepared for scaling

Infrastructure for processing AI content

Preparation for attracting investments