Inspiration

We've all experienced the panic of losing a valuable item and the frustration of navigating disorganized, physical lost-and-found boxes.

What it does

ClankAndFound uses AI to automatically cross-reference lost item reports with found inventory, instantly notifying owners via email when a match is discovered.

How we built it

We built the system with Next.js and Express.js, using Supabase for our database, Google Gemini API for intelligent semantic matching, and Resend for instant email notifications.

Challenges we ran into

Fine-tuning the AI prompts to accurately match vague user descriptions of lost items with specific details of found items without generating false positives.

Accomplishments that we're proud of

Successfully integrating a fully automated loop where a found item submission triggers a real-time AI analysis and immediately dispatches a formatted email alert to the potential owner.

What we learned

We learned how to leverage Generative AI not just for text generation, but as a logic engine to solve the "fuzzy matching" problem in database records.

What's next for ClankAndFound

Implementing image recognition to allow users to snap a photo of a found item on their mobile for even more accurate, visual-based matching.

Built With

Share this project:

Updates