Frigate

Overview

Frigate is an open-source network video recorder for IP cameras that runs local AI object detection on your hardware. It is best suited to camera owners comfortable with self-hosting and Docker who want detection and recording without sending camera feeds outside the home. Its strongest case is local processing paired with fine-grained recording and alert controls; its main trade-offs are host setup and limited native enterprise authentication.

Key features

Frigate can record continuously or retain footage according to detected objects, so users can choose between a complete timeline and storage shaped around events. With a supported AI accelerator, object detection distinguishes people and cars from motion, helping reduce motion-triggered false positives. The maker says compatible hardware can process more than 100 detections per second; that performance depends on having a supported accelerator.

Real-time object tracking and zones let users refine notifications around specific locations, which is useful when alerts should depend on where something appears rather than motion alone. Frigate also supports real-time alerts and behavior and event detection. Integrations include Home Assistant, OpenHab, NodeRed and systems that support MQTT. The Home Assistant integration exposes camera, image, sensor, switch and binary sensor entities, and allows browsing snapshots and recordings.

An HTTP API covers cameras, classification, events, media and recordings, giving other systems a way to work with Frigate's data. User information is stored in Frigate's database, with passwords hashed using PBKDF2-SHA256 at 600,000 iterations. For organizations that require OIDC, SAML or LDAP, Frigate has no native support for those protocols; the documentation points to upstream authentication proxies instead.

Frigate+ adds access to models, up to 12 fine-tuned models and AI-suggested labels for uploads. Model training supports only specific labels, however, and does not yet permit custom labels. That makes the subscription less suitable for teams whose use case depends on a specialized label set.

Pricing

The Frigate plan costs 0.00 USD per free and provides the open-source NVR as a Docker container for a Docker host. It suits users who want the core recording and detection software and are prepared to supply and manage the host themselves.

Frigate+ Subscription costs 50.00 USD per year, billed annually. It adds Frigate+ models, up to 12 fine-tuned models and AI-suggested upload labels, while supporting open-source Frigate. The paid plan is for users who want those model capabilities; it does not remove the custom-label limitation.

Platforms

Frigate supports API, Linux, self-hosted and web environments. It runs in Docker on Docker hosts, including as a Home Assistant App. Documentation recommends Docker Compose and says it runs best with Docker on bare-metal Debian-based distributions. Windows is not officially supported, although some users have run it through WSL or VirtualBox. Windows users should treat that as an unofficial route rather than a supported platform.

Who it's for

Frigate is a strong fit for technically capable home users who want local camera processing, object-aware retention and location-refined notifications, particularly those already using Home Assistant or MQTT-compatible systems. Its API and integrations also suit setups that need to connect camera events to other software. It is a weaker fit for users seeking a turnkey Windows installation or organizations that require native directory or single sign-on protocols.

Pros and cons

• Pro: Local processing keeps camera feeds within the home, according to the maker, which suits privacy-conscious camera owners.
• Pro: Object-based retention, 24/7 recording and zone-refined notifications let users choose how footage is kept and which activity merits attention.
• Pro: Home Assistant, OpenHab, NodeRed, MQTT and a broad HTTP API support integration into existing automation systems.
• Con: Docker is central to installation, and the recommended setup is bare-metal Debian-based Linux, making it a poor match for users who want officially supported Windows hosting.
• Con: Detection above 100 per second and object-versus-motion filtering depend on a supported AI accelerator.
• Con: OIDC, SAML and LDAP require an upstream authentication proxy, while Frigate+ model training cannot use custom labels.

Alternatives

For a broader AI Video Analytics Software shortlist, compare options by deployment and pricing needs. Roboflow is a freemium alternative with Android, iOS, API, self-hosted and web platforms; its free tier includes 10 credits a month, enough to train about 30 models or run 80,000 inferences. It may suit readers focused on model training and inference quotas rather than a local camera NVR.

Amazon Rekognition Content Moderation is an API and web option with a free tier that lasts 12 months and includes 60 free video analysis minutes per month. Consider it when a time-limited hosted video analysis allowance better fits than self-hosting. Camio offers usage-based service levels matched to needs, with pricing available by contacting Camio or an approved partner; it is an option when a matched service level is preferable to managing a Docker host.

Scylla Command Center offers an enterprise plan with 100+ video wall displays, multi-site central management and redundancy with geo-failover, and requires contacting Scylla for pricing. It is more relevant when those enterprise management capabilities are priorities. Lumeo has a free plan and trial, and its Core plan starts at $99/mo with 10 streams, additional streams at $9.99/mo each and 30-day recording retention. Its stream-based plan may suit readers comparing managed stream capacity and retention.

Triya AI is another paid option. Vaidio AI Vision Platform offers a perpetual software license with annual software upgrade and maintenance fees of 15% and annual hardware warranty and maintenance fees of 5%; consider it when a perpetual license is preferred. Videonetics AI Video Analytics has a 30-day free trial and offers a demo, making it an option for readers who want a trial before choosing.

Verdict

Choose Frigate if you want an open-source camera NVR that processes feeds locally and can tie object detection, retention and zone-based alerts into a self-hosted setup. The combination of privacy-focused processing and automation integrations is its clearest advantage. Look elsewhere if you need officially supported Windows hosting, native OIDC, SAML or LDAP, or custom Frigate+ labels.