Snag
Built at the Insforge x Qoder AI Agent Hackathon (Seattle, March 2026), Snag placed 2nd overall out of 30+ competing teams. Snag is an autonomous shopping agent designed as 'Robinhood for your Amazon cart.' Users track desired products using natural language or URLs, and a background AI agent continuously monitors prices across major retailers using quantitative trading signals adapted from finance, including moving averages and volatility analysis. When a product's composite score crosses a configured threshold, the agent sends a real-time buy recommendation, allowing users to review the AI's reasoning and confirm the purchase directly from the mobile app.
Snag's authentication screen — users sign in with email and password or Google OAuth to access their personal AI shopping agent.
The main watchlist dashboard showing five tracked products with live prices, retailer sources, and active deal alerts flagged by the AI agent.
The deals view showing AI-generated buy recommendations — each card displays the current price, savings off target, and buy now or skip actions for the user to confirm.
Product detail view for a tracked item — showing current, target, and all-time-low prices alongside a 20-snapshot price history chart and key pricing statistics.
The activity screen logging confirmed purchases — each entry shows the item name, price paid, and a timestamped confirmation generated by the AI agent.
User profile screen displaying account details, wallet balance, and notification preferences, with a sign-out option.