Never Rely on Memory

Inspiration

After watching a neighbor struggle to file an insurance claim post-storm—missing receipts, vague item details, endless paperwork—I knew there had to be a better way. As a homeowner myself, I’d faced the same stress of tracking household assets, and saw an opportunity to blend AI with simple design to solve this universal problem.

What it does

HomeInventory Manager lets users snap/upload photos of belongings (via mobile/web) – AI auto-identifies items, values, and categories. Users organize items by room/category, edit details (price, receipts, notes), and generate PDF/CSV insurance reports ready to share via email/cloud.

How we built it

We started with core user flows: adding items, filtering inventory, exporting reports. Built the frontend with React/React Native for cross-platform consistency; backend on Node.js with PostgreSQL for data storage. Integrated TensorFlow for AI recognition, Firebase for cloud sync/offline access, and Cloudinary for photo management.

Challenges we ran into

AI recognition accuracy for unique/ low-quality photos was tough—we iterated on model training with real household item datasets to improve detection. Offline sync also posed hurdles; we built local storage logic to queue data for sync once connectivity returned.

Accomplishments that we're proud of

Creating a seamless flow (3 taps max to export reports) loved by beta users; achieving 90%+ AI accuracy for common household items; building full offline functionality that syncs reliably across devices.

What we learned

We deepened our understanding of AI model fine-tuning for real-world use cases, and learned to prioritize offline access—critical for users in areas with spotty internet. User testing also taught us simplicity beats complexity for insurance-focused tools.

What's next for this project

We’ll add barcode/QR scanning for faster item logging, integrate with major insurance providers’ systems, and build a depreciation calculator for asset value tracking. Multi-user access (for families) and push reminders for inventory updates are also in the pipeline.

Built With

• csv-parser
• opencv
• pdfkit
• rest-api
• tensorflow