How Flock Safety cameras work: what they see and store

A plain-English breakdown of the hardware, the data captured, and the networks they plug into.

Published September 25, 2026

Flock Safety is one of several vendors selling automated license plate reader (ALPR) systems to police departments, homeowner associations, and private businesses. This article walks through what the cameras physically do, what data they produce, and where that data ends up.

Hardware

A typical Flock deployment uses solar-powered, pole-mounted cameras with an infrared illuminator, a wide-angle lens, and a cellular uplink. The hardware is purpose-built for plate recognition, not for general surveillance. The image is processed on-device; only the plate string and a coarse location fingerprint are uploaded.

What gets captured

License plate text (the alphanumeric string)
State of issue (where the plate was registered)
Timestamp of the read
Approximate camera location
Vehicle make and model (best-effort, often wrong)
Optional thumbnail of the rear of the vehicle

What is not captured

Driver identity or face
Vehicle speed (in most jurisdictions)
Passenger identity
Audio or video of the interior
GPS coordinates inside the vehicle

Where the data goes

In a typical police deployment, plate reads are streamed into the agency's records management system and into a regional or national hot-list network. Flock operates its own network called FlockOS; competitors run Vigilant LEARN, Axon&s Fusus, and state-level fusion centers. Sharing is governed by the agency's data-sharing agreement with the vendor.

Privacy implications

The combination of plate read + timestamp + location is enough to reconstruct a person's whereabouts in detail. A 2024 Stanford study estimated that a typical urban dweller is captured by ALPR cameras dozens of times per week. Most states have either no retention cap or a rolling 30-day window; a handful cap retention at 90 seconds.