JumboFind

Tufts' smart lost & found hub. Found something? Take a photo, AI auto-fills details & location. Lost something? Check our interactive campus map. Reuniting Jumbos, one item at a time.

Comment

Track: AI

Inspiration

About JumboFind JumboFind is an AI-powered lost and found platform designed specifically for the Tufts University community. We transform the frustrating experience of losing or finding items on campus into a seamless, tech-driven solution. Inspiration The idea came from our own experiences at Tufts. We've all lost things—keys left in the library, jackets forgotten in classrooms, water bottles misplaced at the gym. The current system is broken: items sit in random lost-and-found bins across campus, people spam group chats with descriptions, and most belongings never make it back to their owners. We asked ourselves: what if technology could make this process instant and effortless? What if finding something was as simple as taking a photo?

How we built it

Tech Stack: Frontend: Next.js, React, TypeScript, Tailwind CSS Mapping: Mapbox GL for interactive campus visualization AI Integration: c API for intelligent image analysis and description generation Backend: Next.js Database: SQLite 3 Geolocation: Browser Geolocation API for precise location tracking

Challenges we ran into

  1. Parallel Development Integration: Frontend and backend were built simultaneously by different team members. At integration time, we discovered mismatches in field names (lng vs long), data types, and response structures. We solved this with clear mapping functions and better API documentation.
  2. Simplifying Item Reporting: How do we make logging a found item effortless? Every form field is friction. Our solution: leverage AI to do the work—users just snap a photo, and Claude auto-generates descriptions, suggests tags, and captures location. We reduced a 10-field form to essentially "take photo, submit."
  3. Mapbox Learning Curve: Learning geospatial concepts, coordinate systems, and campus boundaries was challenging. Managing dynamic markers without recreating the map required separating map initialization from marker updates using multiple React useEffect hooks.
  4. Real-time Updates: Getting the map to instantly reflect new backend submissions required careful state management and efficient marker refresh logic.

Built With

Share this project:

Updates