Case study
Grimoire Culinaire
From concept to Google Play private beta, solo, AI-enabled.
The problem
Design an Android mobile recipe app that isn't just a CRUD but a real AI-enabled product: useful daily AI actions, internal ledger economy, gamification, offline-first, 3-tier freemium monetization, Google Play private beta. All as a solo developer, no team, no tech debt.
The method
Full PREDEV scoping before code: product promise, 3-layer Riverpod architecture, breakdown into 14 lots with validation criteria. Each delivered lot goes through tests + post-delivery audit. Internal economy (muffins) structured from lot 1 with reserve-before-debit, automatic rollback and immutable ledger. Multi-LLM routing (OpenAI API primary + Gemini fallback) designed as a product, with quotas, plans and an AI governance surface operable by non-technical admins.
The result
Complete MVP in private beta on Google Play. 400+ tests, 80% coverage. 14 lots shipped without major tech debt, 7 Supabase Edge Functions, 51 analytics events, 18 AI actions with quotas. Riverpod 3-layer + offline-first pattern reusable on other projects. AI governance surface operable by non-technical admins. i18n FR / EN / ES / IT.
Stack
Flutter · Dart · Riverpod 3 · Drift · SQLite · Supabase · PostgreSQL · Firebase · OpenAI API · Gemini · RevenueCat · AdMob