Cameras
Adding cameras, streams, groups, multi-lens devices, and vendor integrations.
Adding a camera
Use Cameras → Discover to scan the LAN for ONVIF devices, or Cameras → Add to enter a device service URL (for example http://192.168.1.50/onvif/device_service) and credentials. LarisVMS then probes the camera: it reads its ONVIF capabilities and streams and picks the Main and Sub streams.
Assign a recorder node, or the camera won't record.
Re-probing
Re-probe on a camera's page reads its capabilities and streams again. Do this after changing the camera's resolution, codec, or firmware. Changing the device URL re-probes automatically on save. Settings → Camera settings can re-probe every camera once a day. Cameras are probed one at a time, so cheap cameras don't drop their recording sessions. Stream names and enabled flags are kept across a re-probe.
Streams
Each camera usually has a Main stream (full resolution, used for recording) and a Sub stream (lower resolution, used for small live tiles and AI detection). Expand Device on the camera's page to rename or disable a stream. Hover a stream to see its RTSP URL.
Groups
Groups organize cameras into sites, buildings, and floors. A top-level group is a site, and groups under it are buildings or floors. A camera can be in any number of groups, as long as they all belong to the same site. Groups are used for filtering, for settings overrides, and for camera access. The built-in All cameras group always contains every camera.
Multi-channel
Some devices have several lenses in one housing. When a probe finds more than one video source, the camera's page offers Split into N cameras. Each lens then becomes its own camera with its own recording and retention. The existing camera becomes channel 1 and keeps its recordings.
Integrations
Some features aren't available over ONVIF. Integrations (listed under Settings → Plugins) are matched automatically from the camera's make and model during a probe. There is nothing to switch on. For example, the Dahua / Amcrest smart events integration reads person and vehicle detections from the camera's own analytics. The camera's Device panel shows when an integration is active.
Also in the app under Help. View source on GitHub.