Back to Catalog
Viseron logo

Viseron

Self-hosted, local-only NVR and AI computer vision software

app

Alternative to: ring, nest, wyze

Repository Docs
Viseron screenshot
About Versions (75)
Back to Versions

v3.7.0

2026-09-17

The main highlight of this release is the new ONVIF component which is a big leap forward for the project. There are some other goodies as well so keep reading.

There is now a dedicated Discord channel for Viseron that I encourage you to join! Here you can ask questions and get help from fellow users. The community is very small at the moment but I hope we can grow it together. You can join at this link

Breaking changes

  • RPi3 support is no longer included in Viseron and new versions (this one included) will not publish the roflcoopter/rpi3-viseron image. I decided to do this because it is an older platform which is very hard to maintain and I believe that not many are using it anymore. If this causes trouble for you, please reach out through an issue on GitHub and we can discuss solutions.
  • The old ptz component has been removed in favor of the new onvif component. You can read more about the ONVIF component further down in these release notes.

New features

ONVIF!

Viseron finally has ONVIF support! A huge thanks to @kaburagisec who has been working on this for a long time. The implementation is quite extensive so I won’t go into detail here, but it allows you to set a multitude of options in a camera’s configuration, basically turning Viseron into a central configuration hub of all your ONVIF cameras. The settings can be controlled directly from the frontend via the Camera Tuning interface, and can be applied without restarting Viseron.

The documentation is quite extensive and I encourage you to read it. You can find it here

PTZ in frontend

With the new ONVIF component comes the possibility to control your PTZ cameras from the frontend as well. Manual control as well as setting/recalling presets is supported.

image

FastALPR

@kilodelta1 added support for running FastALPR for license plate recognition. This is the only component that provides LPR locally in Viseron and does not require an external server like CodeProject.AI does.

You can find the documentation here

Logs in UI

The logs can now be read from the frontend by going to Settings > Logs

You can filter on specific log levels or search for text, as well as tailing the log file live.

image

Recording schedule

You can now set a schedule for recordings. The schedule can be configured independently for event and continuous recordings. Each camera has its own schedule and is configured via cron expressions.

You can find the documentation here

Other notable features

  • New config option require_motion_overlap, contributed by @mvanhorn. Motion overlap helps to reduce false positives by making sure detected objects overlap with a region that also has motion detected. Without it, if a tree blowing in the wind triggers motion in one region, and an object is detected in another region, a recording will start.
  • Docker images are published to GHCR as well as Docker Hub
  • Object detectors and post processors now save a latest_snapshot.jpg, similar to the latest_thumbnail.jpg. The snapshots are exposed over MQTT and to Home Assistant for use in notifications.
  • New PWA configuration
  • Option to disable chown during startup, for faster restarts of the container. You can find the documentation here

Changes

  • Event DB persistence moved away from the main thread, which provides a slight performance improvement in terms of speed and CPU usage.
  • More specific icons for events on the Events page

Fixes

  • Fix YOLO component to run as non-root
  • Resolved over 200 vulnerabilities by updating most NPM and Python dependencies
  • RAM usage reductions by freeing old frames in a smarter way
  • RAM usage reductions by using forkserver for multiprocessing
  • Memory leaks related to camera disconnects fixed
  • Automatically transcode more streams to fix playback issues
  • Numerous security issues fixed related to assigned cameras, path traversal and onboarding

Docker images will be on Docker Hub and GHCR shortly

All changes

New Contributors

Full Changelog: https://github.com/roflcoopter/viseron/compare/v3.6.2…v3.7.0

Changelog on GitHub →