Sèvres is a full-stack web application that enables users to discover fashion through a swipe-based interface while learning their personal style using behavioral data.
The platform dynamically adapts recommendations based on user interactions and generates insights such as aesthetic profiling and style matching.
Features
- Swipe-based fashion discovery (Tinder-style UI)
- Personalized recommendations based on user behavior (likes, swipes)
- Digital wardrobe for saving curated items
- Aesthetic profiling (dominant style detection)
- “Style Twin” matching based on preferences
- Secure authentication system (JWT + OTP + OAuth)
How It Works
- Tracks user interactions such as likes and swipes
- Aggregates category preferences to identify dominant style patterns
- Dynamically updates recommendations based on behavior
- Uses backend APIs to manage user data and session flow
Tech Stack
Frontend:
HTML, CSS, JavaScript
Backend:
Node.js, Express.js
Database:
MongoDB
Authentication:
JWT (JSON Web Tokens), OTP via Twilio API, OAuth (Google/GitHub/Facebook)
Key Backend Features
- Secure user authentication with JWT sessions
- OTP-based login with expiry and retry limits
- RESTful API design for scalable communication
- MongoDB schemas for user and OTP data management
Future Improvements
- Integrate ML-based recommendation system
- Add collaborative filtering for better personalization
- Deploy full-stack app on cloud (AWS/GCP)
Author
Akshita Singh