Sightline
Real-time vision for warehouses that can't afford another manual count.
Inventory accuracy is a labor problem disguised as a data problem. Sightline runs custom-trained vision models on existing camera infrastructure to keep stock counts honest, flag damaged pallets at the dock, and tell you when a slot's been blocked for hours — without adding scanners, beacons, or a procurement cycle.

Use cases
- Continuous cycle counts from existing CCTV feeds
- Damage detection at the receiving dock
- Slot-occupancy and dwell-time monitoring
- Mobile alerts for forklift drivers on wrong-side picks
Technologies
- YOLO / custom CV
- Edge inference (NVIDIA Jetson)
- Kafka
- React Native
- Time-series DB (TimescaleDB)
Industries
- Logistics
- Retail
- Manufacturing
Services
- Custom model training
- Edge deployment
- Alerting & observability
- Mobile companion app