AI wellness app · 2025–26
StayYoung
A live, AI-driven wellness app — owned end to end, from the member experience to Tovi, the in-app AI coach, to the platform behind it.
The problem
Wellness guidance is generic and easy to drop. StayYoung set out to deliver a personalised, conversational experience that meets people in Hebrew, on mobile, and adapts over time — answering their questions in the moment, not after a search. That meant building a real product, not a landing page.
What I built
- Owned the full stack end to end — the member app, Tovi the in-app AI coach, and the subscription, billing, and funnel platform that runs the business side.
- Built Tovi, the AI coach: scoped prompts, streaming Hebrew replies, per-member context persisted across MongoDB + MySQL, and guardrails that keep answers safe and on-brand.
- Designed a Hebrew-first, RTL-correct UI system on React + Tailwind + shadcn that stays fast on a mid-range phone.
- Modelled the data in PostgreSQL with row-level security and SECURITY DEFINER RPCs; auth, OTP, and notification flows run through Supabase Edge Functions.




Impact
- Live on the App Store and Google Play, with real members using it daily in Hebrew.
- Tovi fields hundreds of member questions a week — 24/7, with hard cost and rate guardrails.
- An RTL-first design system keeps the mobile experience fast and consistent.
Stack
ReactNext.jsTypeScriptSupabasePostgreSQL / RLSMongoDBMySQLEdge FunctionsNode.js / ExpressOpenAITailwind