Inspiration

We know losing personal items can be stressful and to it's often very time consuming and complicated to find your lost items, so we decided to build LostLink!

What it does

LostLink is a secure and intelligent lost-and-found platform that helps users recover items quickly while keeping the inventory private. Users submit a photo or description of a lost item, and LostLink automatically cross-references it against a hidden catalog of lost items. Only high-confidence matches are revealed to the user after verification by authorized assistants, preventing false claims and maintaining privacy.

How we built it

We built LostLink using a React-Native frontend for the user interface and a FastAPI backend to handle inquiries and manage the private inventory. Users can submit lost items via photos or descriptions, which are processed by our matching system. A Snowflake database stores items securely, accessible only to authorized assistants. The system uses automated algorithms to generate potential matches, which assistants review before confirming. For added security, we implemented confidence scoring and ownership verification to prevent false claims.

Challenges we ran into

  • Handling image and text matching efficiently to generate accurate potential matches.
  • Preventing fraudulent claims while keeping the verification process smooth for assistants.

Accomplishments that we're proud of

  • Successfully built a secure and functional lost-and-found platform that automates matching.
  • Created an intuitive user interface for submitting inquiries and tracking status.
  • Implemented assistant-only controls for managing the inventory and verifying matches.

What we learned

  • How to combine automation with human verification to ensure reliability.
  • Techniques for secure data handling and privacy-focused design.
  • The importance of user feedback loops in improving system accuracy and trust.

What's next for LostLink

  • Enhance the AI-based image and text matching for better accuracy.
  • Implement real-time notifications for updates on inquiries.
  • Expand features for multi-language support and more advanced fraud prevention.

Built With

Share this project:

Updates