MemoryLens

home
help
care giver dashboard

What inspired us

My grandmother has dementia. Watching her struggle to remember faces of people she loves every day was heartbreaking. We wanted to build something that could help her and millions like her feel less lost.

What we built

MemoryLens is a web app that uses your phone camera to recognize faces and objects in real time. When a family member walks in, the app quietly tells the patient who they are and their relationship. It also listens for conversations and helps the patient remember what was said.

For caregivers, there is a private dashboard where they can add people, objects, and notes. All sensitive memory data is stored on the Midnight blockchain with zero-knowledge proofs, so only authorized people can access it.

How we built it

We used React and TypeScript for the frontend. Face recognition runs locally in the browser using face-api.js. Object detection uses TensorFlow COCO-SSD. We integrated the Midnight blockchain for private, encrypted memory storage with ZK proofs. Supabase handles authentication and backup storage.

What we learned

Building for people with cognitive disabilities taught us that simplicity is everything. Every button, every message, every delay matters. We also learned a lot about zero-knowledge proofs and how powerful privacy-preserving technology can be in healthcare.

Challenges

Getting face recognition to run smoothly in the browser was hard. Integrating Midnight blockchain in a frontend-only app with no backend was a real challenge. We also had to make the UI work for both patients and caregivers with very different needs.