PetsPOS — Multi-Platform Point of Sale
AI-enhanced POS system for US pet businesses — multi-organisation, cross-platform (desktop, web, mobile), with stock analytics and AI-powered reorder predictions.
Problem
Pet store owners in the US were managing inventory manually — frequent stockouts on popular items, overstock on slow movers, and no cross-location visibility for multi-store operators.
Solution
A full POS system covering sales, inventory, and reporting — running on desktop, web, and mobile from a single Flutter codebase. Multi-organisation support for chains. AI-powered stock analytics surface reorder predictions and stockout alerts before they become a problem.
System Architecture
Flutter for cross-platform UI (desktop, web, iOS, Android). Supabase for multi-tenant data, auth, and realtime stock updates. PostgreSQL for transactions, inventory, and organisation data. AI analytics layer for stock trend analysis and reorder predictions.
# system diagram
Desktop · Web · iOS · Android
└─ Flutter (single codebase)
├─ Supabase → multi-org data + auth
├─ PostgreSQL → transactions, inventory
└─ AI analytics → stock predictions + alerts
DevOps & Deployment
Deployment
Supabase (backend) · Web · Desktop · App Store · Play Store
CI/CD
GitHub Actions with platform-specific build targets
Challenges
Building a POS that works reliably offline (desktop) and online (web/mobile) with consistent data sync. Implemented an offline-first transaction queue that syncs when connectivity is restored.
Key Learnings
Flutter for desktop is underrated — the same codebase that runs on mobile runs on Windows and macOS with minimal adaptation. For a POS system, this dramatically reduced development and maintenance cost.