Frigate
Self-hosted camera recorder that spots people, cars and animals locally using AI.
- Stars
- 37.1k
- Forks
- 3.8k
- Licence
- MIT
- Language
- Python
- Last update
- 4 days ago
- Type
- App
What it is
Frigate is a network video recorder (NVR) that connects to IP cameras and runs object detection on the video to identify things like objects moving through the scene. The README says it processes video locally rather than in the cloud, and it is designed to integrate with Home Assistant. It records footage with retention settings based on what was detected.
- Record 24/7 video from IP cameras to local storage
- Get object detection alerts over MQTT for other systems
- Re-stream camera feeds via RTSP to reduce camera connections
- View live camera feeds through a dashboard with WebRTC or MSE
- Edit masks and zones for cameras in the built-in editor
Great for
- People with IP cameras who want local recording and detection
- Home Assistant users looking for camera integration
- Users who want to keep camera footage off third-party cloud services
- People with a GPU or AI accelerator who want faster detection
Not ideal for
- People without any IP cameras, since it needs camera streams to be useful
- Users who want a simple plug-and-play consumer camera system, because the README describes a more technical setup
- People without spare hardware, since the README says a GPU or AI accelerator is highly recommended
How to get it
- 1
Run with Docker
You'll need: Docker
Clone the repository
git clone https://github.com/blakeblackshear/frigate - 2
Run it
- Read the documentation at https://docs.frigate.video
- Choose an object detector suited to your hardware
- Configure your IP camera streams
- Start Frigate and open its web interface to view live feeds
- 3
Get the most out of it
- Add a supported object detector or AI accelerator for better performance, per the README
- Connect to Home Assistant using the custom component
- Use MQTT to send detection events to other systems
- Set retention settings based on detected objects
- Use the built-in mask and zone editor to focus detection
Downloads
Strengths
- Runs object detection locally, as stated in the README
- Integrates with Home Assistant and MQTT
- Open source under the MIT license with an active project
Trade-offs
- The README recommends a GPU or AI accelerator for good performance
- Setup details live in external documentation, so the README alone gives limited install guidance
- Trademarks of Frigate, Inc. are not covered by the MIT license
How it compares
| Project | Stars | Type | Level | Updated | Best for |
|---|---|---|---|---|---|
| frigate (this) | 37.1k | App | intermediate | 4 days ago | People with IP cameras who want local recording and detection |
| RuView | 102k | Other | advanced | 4 days ago | Hobbyists who want presence detection in a room without cameras |
| go2rtc | 14.6k | App | intermediate | 10 days ago | Home users connecting IP cameras to a browser or Home Assistant |
| ProxmoxVE | 30k | Dev tool | beginner | 4 days ago | Homelab users who run Proxmox VE and want quick installs of common self-hosted services |
| systeminformer | 16.4k | App | intermediate | 4 days ago | Windows users who want more detail than Task Manager provides |
| Proxmox | 15.1k | Dev tool | intermediate | 2.0 years ago | Homelab users who run Proxmox VE and want a guided way to create containers or VMs |
Stats synced from GitHub. The assessment above was drafted with AI from the project's README on 2026-10-29 and may miss recent changes — check the README before relying on a command. MIT licence: free to use, modify and distribute, including commercially, as long as you keep the copyright notice.