WhereRU?

Inspiration

“WhereRU?” is the question you ask your missing belongings. We wanted to make finding the answer easier.

On a busy campus, a lost item can end up in a group chat, a classroom, or a lost-and-found desk. WhereRU? brings those disconnected reports together and helps students find useful matches without searching through every listing.

What it does

WhereRU? connects Rutgers students who have lost belongings with students who have found them.

Students upload an item photo, review suggested details, and add a location. The app compares lost and found reports using item attributes, descriptions, location, and time, then sends live in-app alerts when potential matches appear.

A searchable Listings feed lets students filter and shortlist reports. Students can also opt into nearby alerts for deliberately shared search areas. Valuable or sensitive items require private ownership evidence, and both participants confirm a successful return.

How we built it

We used Figma for the design and React, TypeScript, Vite, and Tailwind CSS for the frontend. A Flask backend with SQLite handles accounts, reports, matching, notifications, ownership claims, and return confirmation.

Local image analysis and reviewed reference photos help suggest item details without requiring a Gemini API key. We also integrated a local DINOv2 and OpenCV classroom matcher using Rutgers reference images to help students identify where an item was found.

Challenges we ran into

Two similar items do not necessarily belong to the same person. We needed to suggest useful matches while leaving ownership decisions to the people involved.

We also had to connect the Figma interface to persistent data, keep separate users’ sessions independent, and protect private reports and identifying evidence. Image and room recognition introduced uncertainty, so suggestions remain editable and require confirmation.

Accomplishments that we’re proud of

Connected the owner and finder experience, from posting reports to receiving a match alert and confirming a return.
Added a searchable feed, filters, saved listings, nearby alerts, and automatic seven-day report expiry.
Preserved the Figma design while connecting the interface to the backend.
Passed 118 automated tests and exercised simultaneous users through isolated stress tests.
Built a presentation demo that works without an AI API key.

What we learned

A useful match starts a conversation; it does not establish ownership. Recovery also depends on privacy, clear evidence, and agreement between participants.

We learned to enforce access controls on the server, test users independently, and make uncertain image suggestions easy to correct. Building a working demo also taught us how much needs to happen behind a polished interface.

What’s next for WhereRU?

We want to improve recognition of unfamiliar items, expand the reference collection, and evaluate match quality with student feedback.

Next steps include real email verification, notifications when the app is closed, and a hosted deployment. A campus pilot would help us measure successful returns and understand which suggestions actually save students time.