Frigate
Network video recorder with real-time AI object detection.
Open the official app on frigate.video
This tool is hosted by its maintainers. Click below to open frigate.video in a new tab — it's their official demo.
Browse image & tools tools →What's next with Frigate?
Choose how you want to get started.
Use it free
Open the official tool or demo — no account needed.
Self-host it
Run the open-source version on your own infrastructure.
What is Frigate?
Frigate is an open-source network video recorder (NVR) designed for real-time AI-powered object detection on local hardware. It enables users to monitor IP cameras while processing video feeds entirely on their own devices, ensuring data privacy by preventing cloud transmission. The tool is primarily used by privacy-conscious individuals and home automation enthusiasts who seek advanced security monitoring without reliance on third-party services. Frigate addresses the common issue of false positives in traditional motion detection systems by leveraging local AI accelerators to distinguish between irrelevant motion (e.g., shadows, wind) and meaningful events (e.g., people, vehicles). Its modular design and support for custom models make it adaptable for diverse surveillance needs, from residential security to industrial monitoring.
How it works
Frigate is an open-source NVR that processes IP camera feeds using local AI accelerators for real-time object detection. It replaces traditional motion detection with intelligent analysis, identifying specific objects like people, vehicles, or animals. The tool prioritizes privacy by keeping all video data on local hardware, eliminating the need for cloud storage or processing. It is particularly popular among users who value data sovereignty and want to avoid subscription-based services. Frigate supports real-time object detection at 100+ frames per second using AI accelerators like GPUs or TPUs. It allows users to define detection zones and customize alerts based on specific events, such as a person approaching a door or a vehicle entering a driveway. The tool also integrates with Frigate+ for enhanced model training using real-world camera footage.
How to use it
- 1Install Frigate on a local server or hardware with an AI accelerator (e.g., Raspberry Pi with GPU). 2. Configure camera settings via the web interface, specifying IP addresses and resolution. 3. Deploy pre-trained models or upload custom models for specific objects. 4. Set up detection zones and alert rules through the dashboard. 5. Monitor live feeds and review recorded events with timestamps and object labels. Practical tips: Use a dedicated GPU for optimal performance, regularly update models with user data, and ensure cameras are positioned for clear visibility. Avoid overloading the system with too many cameras or high-resolution streams.
What it can do
- nvr with AI
Use cases
Assumptions and limitations
Assumptions
- source: https://github.com/blakeblackshear/frigate
- license: MIT — free to use
- privacy: Self-hosted — you control your data
Limitations
- Requires dedicated hardware with AI accelerators for optimal performance
- Limited built-in cloud integration for remote access
- Model training demands technical expertise for customizations
- High-resolution video streams may strain system resources
- Depends on camera compatibility with Frigate's protocol support
Understanding the result
Network video recorder with real-time AI object detection.
Tool details
- Clearly flagged when a network request is needed.
- No account, no sign-up, and no tracking of your content.
- Powered by (MIT).
- Built with
- (blakeblackshear/frigate)
- License
- MIT
- Runs locally
- No — requires a network request
- Verification
- Not yet verified
- Input
- Query
- Output
- Text
Built with blakeblackshear/frigate. OpenToolVault provides the discovery and browser interface while crediting the original project maintainers.
- Built with
- License
- MIT
Open-source project
OpenToolVault is an independent directory. We are not affiliated with or endorsed by this project.
References
- / — GitHub Repository
Upstream project · GitHub
- MIT License
Upstream project
Frequently asked
How does Frigate handle false positives compared to traditional NVRs?
Frigate reduces false positives by using local AI accelerators to analyze video frames for specific objects rather than relying on basic motion detection. Traditional NVRs often trigger alerts for non-threatening movements like shadows or wind, whereas Frigate's object detection filters these out, focusing only on relevant events like people or vehicles.
What hardware is required to run Frigate effectively?
Frigate requires a local server or device with an AI accelerator (e.g., GPU, TPU) for real-time processing. While it can run on low-end hardware, performance improves significantly with dedicated hardware. Cameras must support IP streaming and be compatible with Frigate's configuration protocols.
How do I train a custom model for Frigate?
To train a custom model, upload annotated images of the objects you want to detect via the Frigate+ interface. The tool uses these images to fine-tune pre-trained models, improving accuracy for your specific environment. This process may require multiple iterations and feedback to achieve optimal results.
How does Frigate compare to alternatives like RetinaNet or ONVIF?
Frigate differs from RetinaNet by integrating object detection directly into the NVR workflow, whereas RetinaNet is a standalone model. Compared to ONVIF, which is a camera protocol standard, Frigate provides advanced analytics beyond basic video streaming. It combines both AI processing and NVR functionality in a single platform.
What should I do if Frigate fails to detect objects in low-light conditions?
Ensure cameras have infrared or low-light capabilities, and adjust exposure settings in the camera configuration. If issues persist, consider using a higher-quality AI model or adding supplemental lighting. Verify that the camera feed is properly configured in Frigate's settings and that the model is suitable for the environment.