01
Inference on the device
The model runs on-board. Latency does not depend on your network, and neither does uptime — Edge keeps deciding when the link drops.
VRTLST Edge
The camera that thinks for itself
Edge is where the model physically lives. A sealed IP67 device that runs the full VRTLST stack on-board, so inference happens beside the line rather than in a data centre. One PoE cable carries power in and decisions out.
INFERENCE38 ms
THROUGHPUT60 fps
POWER12 W PoE
RATINGIP67 · −20→60°C
What it does
The edge-AI camera & device
02
Nothing leaves the plant
Frames are processed and discarded locally by default. The air-gapped configuration has no outbound connection at all.
03
Managed as a fleet
Adopt a device in Command and it joins the rollout pipeline — staged model updates, health checks and one-click rollback.
How it works
From a camera to a decision
A sealed edge-AI device that runs the full VRTLST stack on-board. Single PoE cable in. Decisions out.
- Full inference on-device — works with the network down
- Zero frames leave the plant unless you say so
- Fleet updates and rollback from Command
- Mount it
- IP67 housing rated −20 to 60 °C, so it goes where a normal camera cannot.
- Run one cable
- A single PoE run carries power and data. No separate inference box in the cabinet.
- Adopt it in Command
- The device registers itself and pulls the model assigned to its group.
- Push models to it
- Roll a new version to one device or two hundred, and roll it back just as fast.
38 ms median on-device inference
60 fps sustained throughput
12 W power over a single PoE run
IP67 sealed, −20 to 60 °C
Specifications
The technical detail
What your integrator will ask for on the first call.
- Inference
- 38 ms median on-device
- Throughput
- 60 fps sustained
- Power
- 12 W over a single PoE run
- Environmental
- IP67, −20 to 60 °C
- Connectivity
- 1 GbE PoE+ · optional Wi-Fi 6 · offline capable
- Management
- staged rollout and rollback from Command
Getting started
How a Edge deployment goes
The same shape every time, whichever product you start with.
- Week 0
One hour of footage
Send us recorded video from a single line. We come back with what the model found in it — before anyone signs anything.
- Week 1–2
One line, one model
We stand up a pilot on the line that hurts most, using the cameras already mounted. Operators start correcting it on day one.
- Week 3–4
Wired into the line
The verdict moves from a dashboard to the PLC, the MES or the alerting channel your shift leads actually watch.
- Month 2
Second line, second site
The model and its thresholds travel. Command stages the rollout so a bad version never reaches more than one line.
- Ongoing
It gets better
Corrections accumulate, drift gets flagged, and new defect classes are a fine-tune rather than a project.
What happens when the network drops?
Nothing, from the line’s point of view. Inference is on-device, so Edge keeps making decisions and queues its telemetry until the link returns.
Can we use it in an air-gapped plant?
Yes, and that is a supported mode rather than a workaround. Models arrive on media, inference runs locally, and the device makes no outbound connection at all.
Do we need a separate inference box?
No. The device is the camera and the compute. One PoE run carries power in and decisions out, which usually removes a cabinet from the design.
How are models updated across a fleet?
Command stages a version to a group, watches the health checks, then promotes or rolls it back. One device or two hundred is the same action.
Keep exploring
The rest of the stack
Command
Every camera, every site, one screen
The control plane for your whole visual estate — deployments, model versions, alerts, access and audit.
Explore Command Vision
One model that understands every camera you own
Detection, segmentation, tracking and classification from a single deployment — no per-camera retraining.
Explore Vision Inspect
Catch the defect before it leaves the line
Sub-millimetre surface inspection at line speed, with a decision the PLC can act on.
Explore Inspect