An AI-powered Lost & Found platform built for college campuses. Uses Google Gemini for semantic matching between lost and found items — goes beyond keyword search to understand context and intent. Built with React, Firebase, and Google Gemini API. Key features: - Model Cascading failover — auto-switches between Gemini, Groq, and OpenAI on rate limits - Dual media strategy — Cloudinary for hosting with Base64-to-Firestore fallback - Client-side image compression via HTML5 Canvas for mobile optimization - Real-time data flow between React frontend and Firebase NoSQL backend National Top 10 — TechSprint AI Hackathon (GDGPCE × Hack2Skill)