ClareTan - styleAI

Dashboard
Login
Home Page
Discover Page
Profile Page
System Architecture Diagram
System Road Map

StyleAI is an AI-powered fashion styling platform designed to help users generate personalised outfit recommendations and visual try-on previews for different occasions. Users can upload selfies, wardrobe pieces, and inspiration images, then receive tailored outfit suggestions based on the event, their style preferences, and practical considerations such as Singapore’s hot and humid climate. The system also supports image-based try-on visualisation so users can preview a look before wearing it. To implement the product, we used React with TanStack Router for the frontend experience and navigation, and React Query for efficient server-state handling and caching. Supabase powers authentication, image storage, and database persistence, allowing uploaded wardrobe items and generated outfit results to be saved and retrieved securely. For AI capabilities, we integrated Reka AI to generate structured outfit recommendations and describe uploaded clothing or inspiration images, while Google Gemini is used for generating outfit visualisation images. User photos are uploaded to Supabase Storage, accessed through signed URLs, and linked to wardrobe and recommendation records in the database.

A key design goal was to make the experience both practical and reliable. Outfit outputs are generated in a structured format so they can be rendered consistently in the UI. Wardrobe uploads are organised into categories such as selfies, clothing, and inspiration, with independent delete controls for easier management. We also reduced unnecessary AI calls by making visual preview generation more intentional and by reusing previously generated images when available.