What is this?
Frigate is a complete, local network video recorder (NVR) that performs realtime AI object detection on IP camera feeds, built specifically for tight integration with Home Assistant.
Frigate is a complete, local network video recorder (NVR) that performs realtime AI object detection on IP camera feeds, built specifically for tight integration with Home Assistant.Is this a standalone app or a Home Assistant add-on?It is a self-hosted server application (run via Docker) that integrates with Home Assistant through a custom component, but it also works fully standalone with its own web dashboard.
It is a self-hosted server application (run via Docker) that integrates with Home Assistant through a custom component, but it also works fully standalone with its own web dashboard.How do I get it running?Frigate is deployed via Docker Compose, pointing it at your RTSP camera streams and configuring it to communicate over MQTT for integration with other systems.
Frigate is deployed via Docker Compose, pointing it at your RTSP camera streams and configuring it to communicate over MQTT for integration with other systems.Does it process every frame from my cameras?No, it deliberately avoids that: it uses lightweight motion detection to figure out where and when to run the heavier object detection step, prioritizing realtime performance over exhaustive frame analysis.
No, it deliberately avoids that: it uses lightweight motion detection to figure out where and when to run the heavier object detection step, prioritizing realtime performance over exhaustive frame analysis.Does my camera data leave my network?No, object detection and recording are performed entirely locally, with no cloud dependency for processing your camera feeds.
No, object detection and recording are performed entirely locally, with no cloud dependency for processing your camera feeds.How does recording and retention work?Frigate records 24/7 and applies retention settings based on whether objects were detected, so footage can be kept selectively around events of interest.
Frigate records 24/7 and applies retention settings based on whether objects were detected, so footage can be kept selectively around events of interest.How do I view the live camera feed with low latency?It supports WebRTC and MSE for low-latency live viewing, and can re-stream cameras via RTSP to reduce the number of direct connections to each camera.
It supports WebRTC and MSE for low-latency live viewing, and can re-stream cameras via RTSP to reduce the number of direct connections to each camera.What features does the web dashboard offer?It includes a live multi-camera dashboard, a streamlined review workflow for events, multi-camera scrubbing, and a built-in mask and zone editor for tuning detection areas.
It includes a live multi-camera dashboard, a streamlined review workflow for events, multi-camera scrubbing, and a built-in mask and zone editor for tuning detection areas.What are similar projects?glasshouse, zigbee2mqtt, local_roborock_server, ha_xiaomi_home, ha-mcp, go2rtc, esphome, mediamtx
glasshouse, zigbee2mqtt, local_roborock_server, ha_xiaomi_home, ha-mcp, go2rtc, esphome, mediamtxWhich topics is it in?rtsp, camera, smart-home, mqtt, home-assistant, home-automation, realtime, object-detection, tensorflow, video-processing, video, computer-vision
rtsp, camera, smart-home, mqtt, home-assistant, home-automation, realtime, object-detection, tensorflow, video-processing, video, computer-vision