FindIt

Inspiration

“Where did I put it?” is a small question that can waste a lot of time. I wanted to make finding everyday belongings easier without buying tracking devices. That idea became FindIt: a searchable visual memory of your spaces.

What it does

Users photograph a drawer, shelf, desk, or storage box and give it a location. FindIt suggests labels for common objects, while users can correct names, add missing items, and mark their positions.

Searching for an item brings up matching photos and its last saved location. Photos and labels stay in the user’s browser, with no account required.

How we built it

I built FindIt as a solo project using React, TypeScript, Vite, and Tailwind CSS. TensorFlow.js and COCO-SSD handle object detection in the browser, while IndexedDB stores photos and labels.

The app runs on GitHub Pages without a backend or paid AI API. I used AI coding assistance and existing UI components during development.

Challenges we ran into

Object detection can miss small, overlapping, or uncommon items. I added manual labels and photo markers so users can still organize belongings the model does not recognize.

The original version also required server storage. Converting it to browser storage simplified deployment, but introduced limits: saved libraries stay on one browser and can disappear when site data is cleared.

Accomplishments that we're proud of

I brought photo scanning, editable labels, object highlights, and search together in a working public app. I also tested saving, editing, deleting, reloading, and failed storage writes to check that the library behaves reliably.

What we learned

I learned how browser-based machine learning and persistent storage work together. More importantly, I learned that useful AI products need clear limitations and ways for users to correct mistakes.

What's next for FindIt

I would like to add library export and import, better organization for larger collections, and improved object recognition. Optional device synchronization could make saved belongings accessible across multiple devices.