Mobile app · building
A private family photo app
Two divorced-and-remarried families, five family group texts, and grandparents on Android while everyone else is on iPhone. I set out to build one private place where each photo reaches exactly the people it's meant for.
The problem
Photos of our kids arrived late, compressed, or not at all, and every send carried the worry of who might feel left out. Existing family apps assume one household, not two equal ones, and we don't post our kids publicly.
What I built
- Groups you name: share with Everyone, one side of the family, or an event like a birthday trip. One person can be in several groups.
- No social pressure: no "Seen by 3 of 8" count. A returned heart shows the photo arrived.
- Built for grandparents: large text and tap targets, words under every icon, and invites that work from far away.
- Privacy underneath: photos are encrypted on the phone before upload, group chat is end-to-end encrypted, a post you can't see behaves as "not found", and photo location data is stripped on the phone. (For technical readers: AES-256-GCM, MLS per RFC 9420, and Postgres row-level security.)
What it looks like
A made-up family with drawn sample pictures. No real people.
How I built it
I ran it as a team of eight AI agents (staff PM, QA, app engineer, UI/UX for grandparents, a UX reviewer for public work, privacy and security, marketing, and funding), routed by a chief-of-staff session. QA writes failing tests first and the engineer makes them pass. There are 600+ app tests and 800+ database checks, and a co-parent data leak was caught in review before any real user saw it.
Stack: Expo / React Native, Supabase (Postgres with row-level security), US-hosted storage.
What's next
A private beta with about a dozen family members, mostly grandparents in West Virginia, targeted for late November.
The demo is a made-up family kept only in your browser: no sign-in, no backend. The code is private; the family beta starts late November.