• 

  System Architecture Diagram

ReUnite – AI-Powered Lost & Found Platform for NTU

Inspiration

The current NTU lost & found system relies on:

• A cluttered Telegram channel (hard to search through long messages)
• A physical Google Sheets at One-Stop SAC (inconvenient for reporting)

We built ReUnite to modernize this process with AI and cloud technology.

Tech Stack

Key Features

Lost Item Reporting

• AI-Powered Descriptions: Gemini generates text descriptions from uploaded images
• Smart Geolocation: Live map preview with drop-off pinning
• Cloud Storage: Images securely stored in Cloudinary
• Auto-Expiry: Items automatically removed after 30 days

Item Search

• Dual Search Modes:
  • Text search in description field
  • Image similarity matching via Gemini
• Flexible Viewing: Grid or list display options
• Detailed Modal: Popup with full item details upon selection

Secure Claim Process

• Google account login required for claims
• Drop-off location only revealed after claiming
• Dispute resolution via campus security

Notification System

• Email subscriptions for unmatched searches
• Daily database checks for subscribed items
• Automatic alerts when matches are found

Admin Dashboard

• Secure login for administrators
• System analytics and reporting
• Full database access and management

Challenges Faced

• Firebase's new credit card requirement forced us to:
  • Evaluate alternative backend solutions
  • Implement careful cost monitoring
  • This added significant development time

Future Roadmap

• NTU Adoption: Seek official university support
• Global Expansion: Scale platform beyond NTU
• Mobile App: Develop native iOS/Android versions
• Enhanced AI: Improve image matching accuracy
• Campus Integration: Connect with security systems

Built With

• css3
• firebase
• gemini
• html5
• javascript
• react