A computer-vision layer that runs on the cameras you already have. Detects hairnets, gloves, aprons, and pests; counts zone dwell time; fires Telegram alerts the moment something is off.
Project
KitchenGuard
Outcomes
8–12
Cameras per box
YOLOv8
Real-time detection
Live
Telegram alerts
1×
Box: mini-PC or Jetson
Kitchen hygiene is the single biggest variable in any F&B operation, and it's checked by humans in periodic walk-throughs. By the time an issue is caught, the shift is over. Brands also need hard data to defend themselves in incidents and audits.
Run YOLOv8 on a mini-PC or Jetson connected to the existing CCTV. Define per-zone rules (prep / line / wash) and per-class rules (hairnet, gloves, apron, handwash events, pests). Stream detections to a Flask dashboard; fire Telegram alerts the moment a rule is violated; aggregate per-shift reports.
A drop-in CV service for any F&B kitchen: existing cameras stay, the box adds 8–12 cameras of real-time detection, and the manager gets Telegram alerts and a per-shift dashboard on port 5000. Built on Python + YOLOv8 + Flask.
Most of our projects start with a 30-minute scoping call.
