# Frigate

NVR with realtime local object detection for IP cameras

Repository: https://github.com/blakeblackshear/frigate
Canonical: https://ross.abutalabs.com/products/frigate
Homepage: https://frigate.video
Language: TypeScript
License: MIT
License Family: permissive
Topics: rtsp, realtime, tensorflow, google-coral, mqtt, nvr, camera, home-assistant, object-detection, ai, homeautomation, home-automation
Last push: 2026-08-26T22:20:37+00:00
Link (homepage): https://frigate.video
Link (site_page): https://docs.frigate.video
Link (site_page): https://frigate.video/plus

## Health v2 (maintenance only)
Score: 92/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 99, release rhythm 78, longevity 100
- inputs: {"age_days": 2776, "days_push": 7, "days_rel": 66, "gap_med": 46.5, "n_releases_24m": 11}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 35407, forks 3533 (observed 2026-08-28T04:12:00.811597+00:00)

## What it is
Frigate is an open-source, self-hosted network video recorder (NVR) that performs real-time AI object detection on IP camera feeds locally using TensorFlow and OpenCV. It integrates tightly with Home Assistant, communicates over MQTT, and supports AI accelerators like Google Coral for high-performance detection.

## Use cases
- monitor security cameras with local AI object detection
- record 24/7 footage from IP cameras with event-based retention
- get alerts when a person or car enters a specific zone
- integrate camera detections into Home Assistant automations
- reduce false positives from motion-triggered recordings
- detect packages or deliveries at the front door
- re-stream RTSP camera feeds with low-latency live view

## When to choose
- you want a privacy-focused NVR where camera footage never leaves your home
- you use Home Assistant or MQTT-based home automation and want detection-driven automations
- you have a Google Coral or GPU and want real-time detection across many cameras
- you need 24/7 recording with retention based on detected objects

## When to avoid
- you only have CPU-only hardware and many high-resolution cameras
- you need a cloud-managed or plug-and-play commercial NVR
- you don't run any home automation platform and just want simple motion recording

## Facets
- artifact type: application
- maturity: active
- function: computer-vision, video-processing, monitoring, self-hosted, machine-learning
- domain: computer-vision, self-hosted, iot, privacy
- platform: self-hosted
- tags: nvr, ip-cameras, rtsp, home-assistant, mqtt, tensorflow, google-coral, security-cameras, object-detection, home-automation, video, docker, linux, web-server

## Member repositories
- blakeblackshear/frigate (main) score 92
- blakeblackshear/frigate-hass-integration (plugin) score 92

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:12:00.811597+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T16:26:56.932361+00:00, confidence not recorded.
  - readme: https://github.com/blakeblackshear/frigate (fetched 2026-08-28T04:12:00.811597+00:00, sha 24f149a9d585)
  - homepage: https://frigate.video (fetched 2026-08-29T07:48:26.057261+00:00, sha 6620a5be9cbf)
  - site_page: https://docs.frigate.video (fetched 2026-08-29T07:48:26.066910+00:00, sha 92bc704779b0)
  - site_page: https://frigate.video/plus (fetched 2026-08-29T07:48:26.069388+00:00, sha 754ff4aa5532)
- Data as of 2026-08-30T08:39:29.467469+00:00.