StyleSwipe

A modern AI-powered virtual try-on application that helps users discover and visualize clothing items that match their personal style preferences. Built with FastAPI, PostgreSQL, and Google's Gemini AI.

What is StyleSwipe?

StyleSwipe is a Tinder-style swiping interface for clothing discovery. Users upload photos of themselves, specify their style preferences, and then swipe through AI-generated images showing them wearing different clothing items. The app uses Google Shopping API to find real products and Google's Gemini AI to generate realistic virtual try-on images.

Key Features

📸 Image Upload & Processing

• Upload three photos (front, side, back views) of yourself
• Supports both JPG and PNG formats
• Automatic image compression and optimization
• Images stored securely in user-specific folders

Style Preferences

• Comprehensive preference form:
  • Gender and size selection
  • Multiple clothing styles (casual, formal, sporty, streetwear, vintage, etc.)
  • Clothing types (tops, bottoms, outerwear, shoes, accessories, etc.)
  • Budget range and color preferences
  • Additional notes for specific requirements

Smart Product Search

• Integrates with Google Shopping API via SerpApi
• Searches UK retailers for products matching user preferences
• Downloads product images and links
• Saves product metadata (price, rating, reviews, source)

AI-Powered Virtual Try-On

• Uses Google Gemini AI to generate realistic try-on images
• Creates combined images from three angles (front, side, back)
• All generated images in 9:16 aspect ratio (perfect for mobile viewing)
• High-quality image generation with natural clothing fit

👆 Tinder-Style Swiping

• Intuitive swipe interface:
  • Tap card → Cycle through front/side/back views
  • Swipe right → Like the item
  • Swipe left → Dislike the item
• Smooth animations and drag interactions
• Progress tracking (X / Y products swiped)

Results & Analytics

• View all liked items in a beautiful results page
• See product details: title, price, source, ratings
• Direct links to purchase products
• All data stored in PostgreSQL for analytics

📊 Grafana Integration

• Full PostgreSQL database for analytics
• Track product clicks, swipe patterns, and user preferences
• Create dashboards for:
  • Click-through rates
  • Popular product types
  • User engagement metrics
  • Conversion funnels

Architecture

Backend

• FastAPI - Modern Python web framework
• PostgreSQL - Relational database for all data
• SQLAlchemy - ORM for database operations
• Google Gemini AI - Image generation and virtual try-on
• SerpApi - Google Shopping product search
• Pillow - Image processing

Frontend

• HTML/CSS/JavaScript - Modern, responsive design
• Dark theme - Beautiful pink/gold gradient accents
• Touch-friendly - Swipe gestures for mobile and desktop
• Loading animations - Smooth user experience

Database Schema

• users - User accounts and metadata
• user_images - Uploaded image paths
• preferences - User style preferences
• products - Products from search API
• swipes - Swipe history (liked/disliked)
• liked_products - Saved liked items
• product_clicks - Click tracking for analytics

Built with ❤️ for HackSheffield

Built With

• frasers
• gemini
• postgresql