#13 · Primary category: Computer Vision

frigate

ai camera google-coral home-assistant home-automation homeautomation mqtt nvr object-detection realtime rtsp tensorflow

NVR with realtime local object detection for IP cameras

Project last updated:08/29/26

GitHub Stars

35.5K

Forks

3.5K

Contributors

400

License

MIT

Why we included this project

Frigate turns a set of IP cameras into a self-hosted NVR that detects people and vehicles in real time, all on your own hardware. It runs object detection locally with OpenCV and TensorFlow, but uses a lightweight motion-detection pass first so the heavier models only run when something actually moves. That keeps CPU and GPU load low enough to watch many cameras at once, and it integrates tightly with Home Assistant through a dedicated component and MQTT. If you want recordings you can search by what happened rather than just by timestamp, this is a practical, well-documented choice for home or small-business setups. It also handles the plumbing you would otherwise build yourself, like 24/7 recording and RTSP re-streaming to limit camera connections, with low-latency live view over WebRTC and MSE.

Articles for this project

No articles for this project yet.

To suggest a topic or contribute an article, contact us.

Related projects in this category