StyleFrame AI

StyleFrame AI brings AI-powered virtual try-on into a retailer's shopping experience. As its primary developer at Twist Digital, I lead technical decisions across the shopper studio, catalogue management, admin and shop-owner portals, backend, and deployment. The retailer platform is implemented; the consumer mobile app remains in development, with its Flutter foundation, API connectivity, and secure token storage established.

I built a QR/barcode-enabled studio that connects a shop's catalogue to a shopper's own photo. Photo validation and multi-garment selection feed an asynchronous generation flow through fal.ai, with progress and result states in the interface. The system manages the work as a job rather than keeping the shopper waiting on a single long request.

A try-on engine also needs tools for the people running the shop. I built catalogue management, admin and shop-owner portals, and usage analytics. Shop-scoped data and role-based access controls separate retailer activity and distinguish platform administration from shop operations.

AI requests consume a real budget. I implemented credit accounting, generation allowances, and spending checks before a job reaches the provider. Session-image cleanup removes temporary assets when a session ends or expires. These controls are part of the generation lifecycle, alongside image validation and handling asynchronous provider results.

The web product uses Next.js, React, NestJS, PostgreSQL, and Prisma. I also established a Flutter and Dart foundation for Android and iOS with MVVM, API connectivity, and secure token storage. Consumer features are still being developed; the mobile app is not presented here as a completed release.