SHIPPED2026
Mobile food-inventory app with photo, receipt, and voice input, Gemini/Claude fallback — live on Google Play, iOS in progress.
Role
Product StrategyUX LogicAI IntegrationMobile DevelopmentSupabase ArchitecturePrompt / Output Design
Stack
ExpoReact NativeTypeScriptSupabaseGeminiClaude
The hard part
Food and home inventory data is messy: users don't want to type everything, images are imperfect, receipts are inconsistent, AI can be wrong, and mobile UX must stay fast and understandable.
A mobile product built around food inventory, AI-assisted analysis, recipes, shopping, and personal utility workflows: a full product system running from camera input through a server-side AI proxy to a personalized experience.
System flow
PhotoReceiptVoiceText
- 01Mobile UXExpo / React Native
- 02Server-side AI proxyJWT + rate limit
- 03Model routing / fallbackGemini → Claude
- 04Structured resultTyped AI actions
- 05Supabase dataPostgres / Storage
- 06Personalized product UIFridge, recipes, chat
Screenshots
Approach
- Camera-first workflows for fridge and receipt capture
- Structured AI actions behind a server-side proxy — no keys on the client
- Model routing with automatic fallback (Gemini primary, Claude reserve)
- Meaningful empty, loading, and error states
- Multilingual product surface across 9 locales
- Android production launch on Google Play
Verification note
Live on Google Play (Android). iOS release pending Apple Developer account setup. No user count or revenue claimed.