Styla

Main Swipe Interface
Saved Items
Recommendation algorithm
Mobile - Swipe Interface
Mobile - Recent Activity

Inspiration

Fashion discovery online is overwhelming—endless scrolling, cluttered feeds, and irrelevant recommendations. We wanted to build something as simple and addictive as swiping, but for style. The inspiration came from Tinder’s clarity of interaction and our frustration with traditional shopping experiences.

What it does

Styla is a fashion discovery platform. Users swipe right to like, left to pass, and up to save. Each swipe becomes a signal that trains a preference model in real time, surfacing items that actually match personal style. Styla also connects swipes to brand sites with affiliate links, enabling instant shopping.

How we built it

Frontend: React + Vite + TypeScript with a swipe-driven UI and gesture animations. Backend: Node.js + Express + MySQL with secure JWT auth, user/item tables, and interaction logging. Data: Ingested brand CSVs (Gymshark, Alo Yoga, etc.), normalized metadata, and optimized indexes.

Challenges we ran into

Engineering swipe gestures to feel fluid across mouse and touch devices. Avoiding stale or repetitive recommendations with diversity bonuses. Scaling ingestion pipelines while maintaining consistent image quality. Balancing fast local learning with server-side synchronization.

Accomplishments that we're proud of

How small interaction design choices (like swipe animations) drastically affect engagement. We are really proud of designing the algorithmic system. Building ingestion tools early makes scaling to multiple brands straightforward.