selfhost.directory

The index of self-hostable projects — ranked, categorized and served with live updates.
Own your data and stop giving it to big tech.

Frigate logo

Self-host Frigate

v0.18.2

Local NVR with real-time AI object detection for IP cameras

ReleaseOpen SourceMIT
Updatedlast month
LanguagePython
Setup difficultyAdvanced
MobilePWA
Stars
37 476
CPU
Intel CPU with AVX + AVX2 instructions (for CPU detection)
Ports
8971 (Authenticated UI and API access without TLS. Reverse proxies should use this port.), 5000 (Internal unauthenticated UI and API access. Intended for use within the docker network for services that integrate with Frigate.), 8554 (RTSP restreaming. By default, these streams are unauthenticated.), 8555 (WebRTC connections for cameras with two-way talk support (TCP and UDP).)
Docker
ghcr.io/blakeblackshear/frigate
Install with Docker
Live view with detection bounding boxesReview items page showing detection eventsMedia browser integration with Home AssistantNotification example with detection snapshot

What is Frigate?

Frigate NVR is a complete, local network video recorder designed for Home Assistant that performs real-time AI object detection for IP cameras. It uses OpenCV and TensorFlow to detect objects locally without sending camera feeds to the cloud. Key capabilities include real-time object detection with support for multiple AI accelerators (Google Coral TPU, Hailo-8, NVIDIA GPUs, Intel OpenVINO, Apple Silicon, AMD ROCm), continuous recording with retention based on detected objects, RTSP restreaming to reduce camera connections, WebRTC and MSE for low-latency live view, MQTT communication for easy integration, and a built-in mask and zone editor. Frigate runs as a Docker container on Linux, with installation options for bare metal, Home Assistant OS, Unraid, Synology, QNAP, Proxmox, and macOS. It requires an MQTT broker (like Mosquitto) and uses SQLite for its database. The web UI supports PWA installation on desktop and mobile. Authentication is built-in with JWT tokens, and it can integrate with external auth proxies like Authelia or Authentik for OIDC/SAML/LDAP. The project is licensed under MIT, though the Frigate name and logo are trademarked.

Should you self-host Frigate?

Frigate scores 83/100 on selfhost.directory (how scoring works) and ranks #1 of 24 in Video Surveillance. It has 37 476 GitHub stars, +969 in the last 30 days.

It shipped 8 stable releases in the past year. The latest, v0.18.2, came out on 15 October 2026. New versions land roughly every 4 weeks.

Setup is advanced: plan for several services and some ongoing maintenance. An official Docker image is available. It is MIT-licensed and written mostly in Python.

Its closest alternative on selfhost.directory is Viseron (69/100, 3 728 stars). Frigate scores higher overall. See Frigate vs Viseron side by side.

Features

  • 24/7 recording with retention
  • Bird classification
  • Face Recognition
  • Home Assistant Integration
  • License plate recognition (LPR)
  • MQTT Integration
  • Multi-camera review
  • Role-based Access Control
  • RTSP restreaming
  • Semantic Search
  • WebRTC & MSE live view
  • Zone and mask editor

Frigate README

logo

Frigate NVR™ - Realtime Object Detection for IP Cameras

License: MIT

How to install Frigate

Install Frigate with Docker

Environment variables
FRIGATE_RTSP_PASSWORD, TZ, FRIGATE_JWT_SECRET

docker run -d
--name frigate
--restart=unless-stopped
--stop-timeout 30
--mount type=tmpfs,target=/tmp/cache,tmpfs-size=1000000000
--device /dev/bus/usb:/dev/bus/usb
--device /dev/dri/renderD128
--shm-size=64m
-v /path/to/your/storage:/media/frigate
-v /path/to/your/config:/config
-v /etc/localtime:/etc/localtime:ro
-e FRIGATE_RTSP_PASSWORD='password'
-p 8971:8971
-p 8554:8554
-p 8555:8555/tcp
-p 8555:8555/udp
ghcr.io/blakeblackshear/frigate:stable

Requires Docker with appropriate hardware device access for acceleration

Frigate releases & version history

@ Every self-hosted release, weekly

One Monday email with the week's stable releases across the directory.

Models for Frigate

Open-weight models you can plug into Frigate, most popular first.

PaddlePaddle logo Detection33MCommercial OK≈0.6 GB↓ 891.6K
Ultralytics logo
YOLO26Ultralytics
Detection?BCommercial OK? GB↓ 27.5K
google logo Detection155MCommercial OK≈0.8 GB↓ 1.5M
Ultralytics logo
YOLO11Ultralytics
Detection?BCommercial OK? GB↓ 32.2K
Ultralytics logo
YOLOv8Ultralytics
Detection?BCommercial OK? GB↓ 35.2K
IDEA-Research logo
grounding-dino-tinyIDEA-Research
Detection172MCommercial OK≈0.9 GB↓ 583K
microsoft logo Detection29MCommercial OK≈0.6 GB↓ 260.5K
PekingU logo Detection43MCommercial OK≈0.6 GB↓ 231.3K

Browse all 25 models →

More in Video Surveillance

Rate this project

Click to rate — no account needed.

0 comments

Posting as guest
Loading comments…