BackToYou

Inspiration

Lost-and-found systems are still manual, slow, and risky. Public item listings expose sensitive details, while staff spend time matching reports by hand. We wanted a system that prioritizes privacy and automation.

What it does

BackToYou allows users to submit lost-item inquiries using text descriptions or photos. These are automatically matched against a private, staff-managed inventory, and information is only revealed when a high-confidence match is found.

How we built it

We built a secure backend that processes text and image inputs, extracts relevant features, and compares them against a protected database. A confidence-based matching layer determines whether a claim is valid before releasing any item details.

Challenges we ran into

We originally wanted to set up face recognition to deter scammers and bad actors but it was more difficult than we anticipated, additionally the phone number verification module was also tricky to implement.

Accomplishments that we're proud of

We delivered a working privacy-first matching pipeline that supports both image and text input, prevents public exposure of inventory data, and meaningfully reduces manual staff effort.

What we learned

learnt a lot about LLMs, database integration with supabase and the intricacies of using and managing multiple server instances.

What's next for BackToYou

Next, we plan to improve matching accuracy, add claim verification workflows, and pilot the system with real-world organizations like campuses, transit systems, and event venues.