Tell your cameras what to watch for.
GenOB is an AI agent that watches your CCTV feeds around the clock. Describe what matters in plain language, like “alert me if someone falls” or “tell me if there’s smoke or fire”, and GenOB alerts you the moment it sees it happen.
Person on the floor
A worker has been lying motionless in Aisle B for 12 seconds.
Guards monitor dozens of feeds at once. Attention drops within minutes, and most incidents are only found later on playback.
Line-crossing and motion zones flood you with false alarms and can’t tell a fall from someone tying their shoe.
A vision-language model reads each camera the way a person would, so you can ask for exactly what you care about.
From camera feed to alert in three steps.
No retraining, no drawing zones, no rule engines. If you can describe it, GenOB can watch for it.
Plug in your existing cameras
Connect GenOB to the CCTV you already have, via RTSP streams or your VMS. Your footage can stay on your own premises.
Lobby · Car park · Warehouse · Wards
Tell it what to look for
Write watch instructions in everyday language and assign them to any camera, zone or time of day.
Get alerts that explain themselves
When something matches, you get an instant alert with the clip, a snapshot and a short description of what happened.
“One person entered using a propped-open door.”
One agent, any instruction.
These are just starting points. Write your own for your site, your risks and your operating hours.
Alert me if a resident falls or is lying on the floor.
Tell me immediately if you see smoke or flames anywhere.
Notify me if anyone climbs the fence or enters the yard after hours.
Flag any bag left unattended for more than two minutes.
Alert me if someone on the loading bay isn’t wearing a hard hat.
Let me know if a queue forms past the yellow line or a crowd gathers.
Tell me if a fire exit is blocked or a door is left propped open.
Alert me if you see a fight or someone acting aggressively.
Notify me if a vehicle stops in the fire lane for over a minute.
Not another motion detector.
Traditional video analytics only recognise what they were trained for. GenOB runs on Video Language Models, so it understands context, actions and intent.
Traditional video analytics rules
- Fixed list of detectors: person, car, line crossed
- New use case means new model, new training data, new project
- Triggers on shadows, rain and passing headlights
- Alerts say “motion in zone 3” and nothing more
GenOB video language model
- Watches for anything you can describe in a sentence
- Add or change instructions in seconds, no retraining
- Understands context, so it knows a fall from a person sitting down
- Every alert comes with a plain-language explanation and the clip
Fits into your security setup.
GenOB is a standalone engine that works alongside your cameras, VMS and control room, not a replacement for them.
On-prem, edge or hybrid
Run on your own hardware so footage never leaves your site, or mix edge and central processing.
Alerts where you work
Push alerts to your control room, VMS, email, mobile or existing incident systems.
Search past footage
Ask questions about recorded video, like “when was the red van last at gate 2?”
Learns from feedback
Mark alerts as correct or false and GenOB gets sharper at what matters on your site.
Take GenOB to your next meeting.
Download the two-page GenOB brochure for an overview of the engine, how it works and its key features. Scan the QR code to open it on your phone.
Scan to view the brochureSee GenOB on your own cameras.
Tell us about your site and what you need to watch for. Our engineers will scope the right setup, whether on-premise, hybrid or edge.
Distributor
Oneberry Technologies Pte Ltd150 Sims Drive, #06-00 Oneberry
Singapore 387381