Inventory AI

InventoryAI is a finished, working Android app that keeps stock counts for small shops, resellers and micro-warehouses - no clipboards. Point the camera at a shelf and the AI counts what is there; scan a barcode to log an item in seconds; when stock runs low, the app builds a restock plan and compares supplier quotes by true landed cost.

WHAT WORKS TODAY (all verified on a real device or in CI):

- Core inventory, offline-first local database

- ML Kit barcode scanning - device-verified

- Photo shelf scan with confidence scores; scans land on a review screen first - nothing reaches stock until the user taps Save

- Confirm-then-save counts that set stock by exact name match

- Opt-in auto-pilot for high-confidence counts (off by default, full history + undo)

- Restock planning from per-item par levels and lead times

- Supplier quote comparison by landed cost, draft purchase orders

- eBay connect screen (publishing built, not yet proven end-to-end)

- User accounts (AWS Cognito) and a LIVE AWS backend (API Gateway + Lambda, EU region), per-user AI quotas enforced server-side

- 434 automated tests, full CI green; 33,000+ lines of code

HONEST STATUS: pre-revenue, zero users, no Play Store listing yet. A video shelf-scan crash was root-caused on-device and the fix passes full CI; the final on-device retest is the one step still open. Subscription billing is designed and documented but not switched on.

HOW IT MAKES MONEY: a per-business subscription - Solo $20/mo (1 device), Team $49/mo (up to 5), Business $99/mo (up to 15). Every tier caps AI spend below its price (Solo caps at $15), so a maxed-out subscriber still leaves margin. Scenario math with average usage at 30% of cap, after a 15% store fee: 25 subscribers = ~$420/mo net; 100 = ~$1,945/mo; 300 = ~$6,195/mo.

WHY I AM SELLING: I am a solo builder and I have taken it as far as I can - the remaining steps (Play Store launch, billing, m