EzVision platform
Book a 20-min call
The EzVision platform is on-premise video analytics software that joins CCTV, plate cameras, thermal cameras, radar and IoT sensors into one system.
It notices what matters, decides who needs to know and acts, across every site you run, on servers at your own premises.
Four steps run on every frame, so you get one event instead of five alerts.
AI models find people, vehicles, objects, behaviour and hazards on every frame.
Camera, radar and sensor signals about the same event become one event.
Your rules decide whether it matters: which zone, what time, how long, how serious.
The right person, screen or system gets it: phone, control room, gate, PLC or API.
Every alert is a rule: zone, object, condition, schedule and action, set on a map of your site.
If a person enters the back fence zone between 22:00 and 06:00, alert the guard on duty.
If a forklift comes within 2 m of a person in Aisle 4, log a near miss and alert the shift lead.
If water rises above 10 cm in Basement B2, alert facilities and close the ramp barrier.
A camera sees a car; your toll system says what it paid. EzVision joins the two, so you hear about the exception.
Vehicle and plate at a lane, checked against toll, tag or payment records.
Bay occupancy checked against bookings and charging sessions.
People through a door, checked against access-control swipes.
Pallets and trailers checked against the WMS and dock schedule.
Aircraft on stand checked against flight and gate status.
Checkout activity checked against POS transactions.
Standard IP, plate and thermal cameras, plus sensors for places video can’t go or shouldn’t.

RTSP and ONVIF streams from the cameras you own.

Plates at gates, lanes and barriers.

Darkness, smoke and rain.

Presence and falls, with no image captured.

Washrooms and stairwells.

Basements, ramps and drains.

Doors forced or held open.

Cold rooms and plant rooms.
A short clip, the place and one tap to acknowledge, on the screen each person uses.

For guards, nurses, teachers and supervisors: a short clip, the place, one-tap acknowledge.

For operators: prioritised alerts on the video wall, with the clip ready to play.

For managers: trends, response times and compliance by site.
Send events to your VMS, access control, gates, WMS, nurse call or PLC.
POST /webhooks/ezvision
{
"event": "intrusion",
"zone": "Back fence",
"time": "22:14",
"clip": "clip.mp4",
"notify": "guard_on_duty"
}Search by plate, vehicle colour, clothing, time and place across every camera, or type what you’re looking for.
Detection runs on GPU servers at your premises, with up to 50 cameras per server.
| Option | Best for |
|---|---|
| On-site server | Footage that must stay on premises; low bandwidth sites. |
| Hybrid | On-site detection with a remote dashboard and search across sites. |
| Multi-site | A server at each site and one view for head office. |
Per camera or per site, depending on the detections you run and how it’s deployed.

Cameras, network and the events that matter.
Existing streams in; server sized.
Alerts checked, not sent, while rules are tuned.
Alerts go to your team on one site.
Same rules, every site.
We train new detections on footage from your own site, then add them to your library.

We collect footage of the event from your own cameras.

We label it and train a model for the event you care about.

We test it on your cameras and add it to your library.
Search the full library, or filter it by industry and category.
Don’t see what you need? We train custom detections on your own footage. Ask about a custom detection
Footage, plates and faces stay under your control.

Usually not. EzVision runs on most standard IP cameras over RTSP or ONVIF, and a site visit confirms what yours can support.
On GPU servers at your site. Footage does not need to leave your premises for detection to run.
Up to 50 cameras per server, depending on the detections running. We size it during the site assessment.
Yes. Events go to your VMS, access control, gates, WMS or PLC through the API, webhooks and standard integrations.
From connected cameras to live alerts in weeks, not months. A shadow run tunes the rules before alerts reach your team.
Per camera or per site, depending on the detections you run and how it is deployed.
Yes. We train custom detections on footage from your own site and add them to your library.
Book a slot or send a quick brief. A real person on our team replies within one working day.
A focused call on your goals, systems and site. We tell you straight what will work, and what won't.
A clear written plan shaped around your needs, with a fixed price and timeline. No surprises.