Production lines were tracked with pen & paper.
Yields were unclear. Miscounts and oversupply went unnoticed.
What if every seedling was automatically scanned and counted?
Real-time data is streamed from barcode scanners and laser counters.
iPads surface issues instantly. Accurate metrics now, not next week.
Line 1
NOW PRODUCING: Alyssum White
scroll to explore
▿
Architecture
- Headless Raspberry Pis read USB barcode scanners and GPIO laser counters
- Node.js + Express REST API aggregates multi-sensor data from all production lines
- Server-Sent Events push state to React iPad dashboards
- File-backed persistence with in-memory index for sub-ms query times
Real-time operations
- SSE pushes every scan and count to iPads instantly
- UI built in collaboration with production line staff
- One-tap on-the-fly adjustments with full audit trail
- Supervisors, managers and accounting all get custom live dashboards
Audit writes
atomic ✓
flush
Network queue
retry OK
queuing
Index verify
healed
mismatch
Scan fallback
degraded OK
bad scan
Built for the floor
- All events queued and written atomically to audit-ready files
- Network drops handled seamlessly by queue and retry with backoff
- In-memory index auto-verifies and self-heals from single source of truth
- Production never stops: graceful degradation with bad scans and sensor faults