Skip to main content
A single device is easy. A fleet — dozens of cameras across lines and sites — is where AcuSight earns its keep. This tutorial takes you from an empty Devices page to a grouped fleet running the same model, managed entirely from the dashboard. You’ll connect devices, organize them into a group, deploy a model fleet-wide, and watch their health on live dashboards — without touching the machines directly after setup.

​
Before you start

You’ll need:
  • An AcuSight account.
  • A trained model to deploy. If you don’t have one, walk the Build a defect detector tutorial first.
  • One or more Linux edge machines (arm64, amd64, or arm) with Docker and SSH access — or virtual devices if you want to try the flow without hardware.

​
Step 1 — Connect your first device

From Devices, click Add Device to generate a one-time setup command, then run it on the edge machine over SSH:
The installer pulls the AcuSight agent, starts it, and registers the device against your organization. It connects outbound over HTTPS, so devices behind NAT or a firewall work with no open ports.
Devices page with a list of connected devices

Each device registers itself and appears on the Devices page with live status.

Full walkthrough, including virtual devices: Connect a device.

​
Step 2 — Add the rest of the fleet

Repeat Step 1 for each device. Setup commands are single-use and expire after 24 hours, so generate a fresh one per machine. Devices come online independently and appear on Devices as they register — there’s nothing to refresh or poll.

​
Step 3 — Organize into an edge group

Managing devices one by one doesn’t scale. Scroll down the left menu to find Edge Groups. Click Add Edge Group to define a named set — for example, Line 3 cameras — and treat it as a single target.
Edge Groups with a named group of devices

Group devices by line, site, or model so you can act on them together.

A device can belong to more than one group, so you can slice the same fleet by line and by site. See Fleet management for the full set of controls.

​
Step 4 — Deploy a model across the group

Open Models, select your model, and choose Deploy, then pick the group — every device in it picks up the model on its next check-in. Rolling an update later is the same action: deploy the new model to the group and the whole fleet moves together. More in Deploy a model.

​
Step 5 — Monitor the fleet

With devices running, keep an eye on them from Monitoring — fleet and model health on live dashboards, so you catch an offline device or a drop in detections early.
Fleet monitoring dashboard

Fleet and model health on live dashboards.

See Monitoring for the dashboards behind these metrics, and Reliability for how devices recover on their own.
If a device shows as waiting or offline, confirm it has outbound internet access and re-run its setup command. Devices reconnect automatically once they can reach AcuSight.

​
Next steps

Fleet management

The full set of per-device and group controls.

Supported devices

Hardware that runs the AcuSight edge agent.