
Collectible Scanner - AI Price Engine
Native iOS app (Swift/SwiftUI) combining camera AI, voice assistant, and multi-source price aggregation for real-time retail arbitrage. Dwell detection auto-scans items, Gemini 2.0 Flash identifies collectibles, and the arbitrage engine pulls prices from PokemonTCG, PriceCharting, and eBay APIs to generate Buy/Watch/Pass recommendations.
AI-Powered Retail Arbitrage Tool
Built a native iOS application that enables collectors and resellers to instantly identify and price collectibles using their phone camera. The app combines computer vision, voice AI, barcode scanning, and multi-source price aggregation into a seamless scanning experience.
Core Features
- Dwell Detection: Hold camera steady and the app auto-scans using CoreMotion sensors
- AI Visual ID: Gemini 2.0 Flash REST API identifies collectibles from camera frames
- Voice Assistant: Gemini Live WebSocket for hands-free queries (“what's this worth?”)
- Barcode UPC Fallback: Scan barcodes when visual ID isn't sufficient
- GPS Store Detection: Apple MapKit locates nearby stores for price comparison
Price Aggregation Engine
- PokemonTCG API for trading card prices (free tier)
- PriceCharting API for retro games and collectibles
- eBay Browse API with OAuth2 for real-time market data
- Arbitrage recommendations: Buy/Watch/Pass based on price spread analysis
- Watchlist with price drop alerts for tracked items
Technical Highlights: Built with Swift 5.9+ and iOS 17+ @Observable pattern. Camera-first dark-themed UI with AVFoundation for async video frame processing. Supabase backend for user data and watchlists.