Git Branch - Elytsx

System Architecture Diagram

Inspiration

Most people own more clothes than they realise yet still feel like they have nothing to wear. We wanted to close that gap and reduce impulse buying by building a personal stylist that lives in your phone.

What it does

Elytsx is an AI-powered wardrobe app for digitising, organising, and styling your clothes.

Digital wardrobe — photograph a piece; AI tags it with category, occasion, weather suitability, and colours automatically
Taste profile — a style quiz on first launch calibrates recommendations to your aesthetic
OOTD — two modes: fully AI-generated, or guided (feed in weather, occasion, vibe, or a specific piece to build around)
Studio — manually build and save outfits; browse history in a collage-style grid
Calendar — attach outfits to events and track what you wore each day

How we built it

Layer	Tech
Frontend	Flutter
Backend	Supabase (Edge Functions + PostgreSQL + Storage)
AI	Reka AI (multimodal vision)

Key decisions:

Repository pattern (WardrobeRepository, OutfitRepository, TagsRepository) keeps data fetching separate from UI
Supabase Edge Functions handle save-wardrobe, get-wardrobe, create-outfit, get-outfits with cursor-based pagination
Shared WardrobeAddState flows through the multi-step add-item flow (upload → AI tagging → review)

Accomplishments that we're proud of

End-to-end AI clothing classification — photo in, structured tags out via Reka
ClothingTag widget system with colour swatches and icon fallbacks across all screens
Full outfit creation flow (manual + AI) backed by real Edge Functions