Occivar is building the intelligence backbone for smart cities and buildings, privacy-first edge AI that counts people, measures density and surfaces anomalies in real time. No video stored, no video transmitted, no faces recognised. Only anonymous numbers ever leave the device.
Three things, in the right order. The sensor sees the room without seeing the person. The platform understands the pattern. The integration acts - adjusting HVAC, flagging anomalies, or simply giving you the data you've been guessing at.
Small ceiling- or wall-mounted sensors run the model on-device. Raw pixels never leave the room. What leaves the sensor is a count, a heat-map, a queue length - anonymous numbers, nothing more. Meet the occupancy sensor.
The platform turns counts into context - which zone is busy, when CO₂ is creeping up, where the queues form. You finally see how the building is actually used, not how the booking system says it should be.
Out to your BMS, Slack, Teams, Snowflake or a webhook - schedule HVAC against real usage, alert when a threshold is crossed, send the right report to the right person on Monday morning.
Real-time counts across every zone. No faces. No video stored. Just the room, abstracted into anonymous dots and densities. Hover a zone for detail.
Demo data simulated locally. Real deployments stream from edge sensors over MQTT or HTTPS.
"No faces. No video stored. Just counts." - what every sensor on this floorplan promises, encoded in the architecture itself.
We don't lead with surveillance, and we don't ask you to trust a policy. The architecture is the promise. The numbers are the answer. The integration is what you do with them.
We measure ourselves in tonnes of CO₂ avoided, dollars not spent on over-cooling, and minutes shaved off queues - not in screens shipped. Every conversation starts with the outcome and works backwards.
Raw video never leaves the sensor. The model runs on-device, emits anonymous counts, and forgets the frame. There is nothing to leak, nothing to subpoena, nothing to misuse - because nothing is stored.
Sub-second counts, calibrated per deployment, with confidence intervals you can pull into a report. Independent ground-truth audits supported on every pilot.
Heatmaps, dwell curves, density-over-time - designed so a Head of Property can read them in 30 seconds and an executive can read them in 5. Built for the meeting, not the analyst.
One sensor in a single room is a useful pilot. A hundred sensors across a campus is the same software. The platform is hardware-agnostic, BMS-friendly and built for organisations that move slowly on procurement but quickly on outcomes.
Our pricing is per zone, not per camera or per face. You pay for the questions we answer, not the data we hoover up. We don't have the data to hoover.
Start a conversationOn the left, the surveillance era - faces tagged, identities tracked, footage hoarded. On the right, Occivar - anonymous dots, counts only, never a frame stored. Arrow keys work too.
We sell the same platform to a council, a vice-chancellor and a retail GM - because the privacy model lets all three say yes.
Libraries, study spaces, lecture theatres.
Civic squares, libraries, town halls.
Footfall, dwell, conversion zones.
Anonymous safety monitoring.
Stations, concourses, platforms.
Crowd flow, density safety, exits.
Office utilisation, hybrid attendance.
Site safety, headcount, zone access.
A worked example of what a deployment would report, sample data, simulated locally. Counts animate on scroll. We are not yet running field deployments.
Demo data for illustration only. Field deployments targeted H2 2026.
Move the sliders. Right-sizing HVAC against real occupancy typically returns 8–18% on a commercial energy bill, plus a proportional emissions reduction. These are conservative defaults - your mileage will vary by climate, age of plant, and how badly your current system is over-provisioned.
Annual saving = floor area × spend per m² × % reduction. Emissions avoided uses 0.68 t CO₂e per MWh - the Australian NEM grid factor - and a 30 c/kWh wholesale-to-retail average.
Conservative assumptions on purpose. Pilots routinely beat the defaults because most commercial buildings are running HVAC to a schedule designed for the busiest day they will ever have.
Walk us through your buildingActive engagements and grant support. Field deployments targeted H2 2026.
"How was the library used last Friday?" · "Why is HVAC running at full tilt on a quiet floor?" · "Are we anywhere near our crowd-safety threshold tonight?"
