Local NVR for IP cameras with realtime object detection and Home Assistant integration
MIT
- TypeScript
- Python
- Shell

About Frigate
Frigate is a local NVR for IP cameras with AI object detection. It is designed for Home Assistant and keeps detection on your own hardware, with OpenCV and TensorFlow used for realtime analysis. The focus is on recording and reviewing camera events without sending video to a cloud service.
It uses low-overhead motion detection to decide where object detection should run, and it can use a GPU or AI accelerator. It supports RTSP re-streaming to reduce camera connections, 24/7 recording, retention settings based on detected objects, and low-latency live view through WebRTC and MSE. It also communicates over MQTT for integration with other systems.
Frigate uses a custom Home Assistant component and supports multiple object detectors. The code is MIT licensed, while the Frigate name, Frigate NVR brand, and logo are trademarks of Frigate, Inc. It is maintained as a self-hosted application for local deployment.
Key features
- Realtime local object detection for IP cameras
- 24/7 recording with object-based retention
- RTSP re-streaming to reduce camera connections
- WebRTC and MSE support for low-latency live view
- MQTT integration and Home Assistant custom component
Details
- First released
- 2019
- Cameras
- IP · RTSP
- Detection
- AI objects · OpenCV · TensorFlow
- Hardware
- GPU · Coral · AI accelerators
- Deployment
- Self-hosted · Docker
- Integrations
- Home Assistant · MQTT
