Back to portfolio
Creator (solo)
Arisium
iOS and Android fitness and nutrition app: AI photo-based meal estimation, workout tracking across a large exercise library and body progression. Designed, coded and shipped solo, from design to billing.
ExpoReact NativeSupabaseAI
- Year
- 2026
- Status
- Private
01The context
A personal product: bring nutrition tracking, training and body progression into a single mobile app, removing the friction of food logging with AI. Built for daily use, with a real subscription funnel.
02My role
Product design and full development, solo: cross-platform mobile app, database, around a dozen AI edge functions, payments and subscription funnel. From authentication to the conversational coach, including photo vision and in-app billing.
03Key features
- AI macro estimation from a photo via Claude Sonnet, in seconds, with a confidence level (high, medium, low)
- Nutrition journal and food search on the French CIQUAL database, complemented by OpenFoodFacts
- Workout tracking: several hundred exercises (French-translated library), automatic PR detection and 1RM estimation (Epley formula)
- Body progression: measurements, before/after photos, charts and goal projection via linear regression
- Conversational AI coach (nutrition and training) and generation of workout programs and batch-cooking plans
- RevenueCat subscriptions: two plans, in-app purchases on iOS and Android, sync webhook
04Tech stack
Expo SDK 55 + Expo RouterReact Native + NativeWind (Tailwind)Supabase (Postgres, Auth, Storage, RLS)Deno Edge FunctionsClaude (Sonnet 4.5 vision, Haiku 4.5 coach)RevenueCat (in-app purchases)EAS Build
05Highlights
- Seven product phases shipped, from auth to AI coach
- Complete fullstack mobile architecture, designed and coded solo
- Multimodal AI: photo vision and conversational coach
- Around a dozen AI edge functions (estimation, coach, program and recipe generation)