AIdataNewsGlobal General Desk: A Public Lookup Shows How Far Flock Camera Coverage Reaches in the US
Figuring out which automated reading systems sit on your own street used to mean filing paperwork with a local agency and then waiting. The search-by-ZIP-code interface now circulating in coverage of public safety technology compresses that into one query: it plots device positions, summarizes the fields each capture produces, and credits the independent research behind the inventory. For readers following how data-collection infrastructure quietly scales, the tool works less as a novelty than as a documentation aid — and it also makes the limits of public visibility explicit, because a private firm controls the underlying records while details about law enforcement use can remain unclear.
The Numbers That Frame the Story
- Scale of collection: the vendor says its network processes more than 20 billion vehicle reads a month across 49 states.
- Devices listed in the inventory: more than 170,000 cameras and over 130,000 other Flock devices.
- Gap against the published figure: Flock has publicly cited around 120,000 cameras, so the camera tally in the inventory is about 40% higher.
- Vintage of the data: a snapshot from December 2025.
- Origin of the dataset: assembled through independent research by cybersecurity researcher Joshua Michael, who identified an exposed access token in November 2025, reported it, and published his findings and his own map in December.
- Accountability record: a Washington Post review found at least 50 officers accused or charged with misusing license plate readers, often to follow people in their personal lives; Flock's system was involved in 46 of those cases.
What a Single Pass Produces
A Flock camera photographs each vehicle that passes it. Suspicion is not a precondition: the unit does not wait for a tip, a warrant or a BOLO description before it writes an image. The base record pairs the plate with a time and a location, which is enough to establish whether a particular car was on a particular road at a particular minute.
The richer layer is the vendor's Vehicle Fingerprint feature. Alongside the plate reading it logs make, model and color, plus surface details such as dents and bumper stickers, so a car can be found even without a readable plate. That combination is what turns the archive into a search instrument: the visual outline of the vehicle functions as an identifier in its own right.
Storage, Search Rights and Cross-Border Sharing
Those records move into Flock's cloud rather than staying on a local machine. From there police departments can search them and share them with other agencies across city and state lines. The design is efficient for investigations and hard to supervise at the same time: one query in a single jurisdiction can surface movements observed in dozens of others, while the record of who queried what sits with the vendor.
Broad access leaves room for misuse and abuse, and the reporting summarized above is the clearest published indicator of that risk. Questions put to the company's chief executive, Garrett Langley, have centered on two points: how the company responds when a user abuses the platform, and who controls the data its devices collect.
How the Device Inventory Was Assembled
The locations shown on the map trace back to independent research, not to a product release. In November 2025, Joshua Michael found that Flock's servers publicly exposed an access token anyone could obtain without logging in. He reported the issue to Flock and then published his findings, together with his own map, in December.
His inventory lists more than 170,000 cameras and over 130,000 other Flock devices. Because Flock has publicly cited around 120,000 cameras, the actual camera count is about 40% higher than the number most often repeated in coverage of the product. The map displaying these results does not host or distribute the underlying dataset; it mirrors it.
Vendor Figure vs. Researcher Inventory
| Measure | Figure cited publicly by the vendor | Figure in the researcher's inventory | Reading the difference |
|---|---|---|---|
| Cameras | around 120,000 cameras | more than 170,000 cameras | The inventory is about 40% higher |
| Non-camera hardware | Not part of the public camera citation | over 130,000 other Flock devices | Enumerated separately by the research |
| Read volume | more than 20 billion vehicle reads a month across 49 states | Not restated by the inventory | A vendor-stated operating metric |
| Currency of the data | Undated in the public citation | December 2025 | Point-in-time snapshot, not a live feed |
Who Does What in This Episode
- Flock Safety — the private company that builds the automated license plate readers, operates the cloud store, and is the named point of contact for the reported exposure.
- Joshua Michael — the cybersecurity researcher whose discovery of an unauthenticated access token produced the device inventory now displayed.
- Garrett Langley — the company's CEO, addressed on misuse response and data control.
- Police departments and other customers — the parties that search the records and pass them to agencies outside their own city or state.
- Residents and municipal leaders — the people who can use the map as a starting point for a local conversation about deployment.
Practical Questions Before You Open the Map
What is the product being plotted?
A Flock camera is an automated license plate reader made by Flock Safety. It photographs passing vehicles, reads their plates, and logs where and when each car was seen. Police departments and other customers can search those records.
How do I check my own area?
Type a ZIP code into the map's search bar. The view zooms to that area and shows every camera in the dataset there; you can also zoom and pan manually to explore any neighborhood in the US.
Why does this inventory exist at all?
It is the by-product of a security disclosure. The displayed data mirrors an inventory of Flock devices that Joshua Michael retrieved from Flock's own systems through an access token the company had left exposed, a condition he located in November 2025.
Is this how volunteer-built maps work?
No. Maps like DeFlock rely on volunteers reporting cameras they spot. This inventory comes from Flock's records instead, which is why its totals run higher than a crowdsourced list would typically capture.
Does the map carry the dataset itself?
It does not. The interface presents the device locations assembled by the researcher's independent work, and does not host or distribute the underlying dataset.
Can I treat the coverage picture as complete?
You cannot. The data is a snapshot from December 2025, so devices installed or removed after that date will not be reflected.
What should I do if I want more local clarity?
Use the map to identify candidate devices, then take the question to municipal leaders: because a private company controls the data, and details about law enforcement use can be unclear, the tool supports informed movement through local areas and possible follow-up rather than a definitive answer.
What to Watch Next
The most useful thing this tool does is reframe a debate that usually runs on trust. Instead of asking whether a reading network is large, it lets anyone count the hardware street by street, and in doing so exposes a second-order question: the vendor's own public number sits well below the enumerated fleet, while the operating metric of more than 20 billion vehicle reads a month across 49 states describes throughput, not coverage. Whether published counts, access controls, or the audit trail around officer misuse shift in response is the open thread to follow.
For now, the practical takeaway is procedural. A ZIP code lookup, a device inventory compiled by Joshua Michael, and a Washington Post finding of at least 50 officers accused or charged with misuse — with Flock's system involved in 46 of those cases — together describe an information environment where visibility arrives from disclosure rather than from catalog copy. Readers who want to know what watches their commute now have a way to start asking, and a documented reason to keep asking who controls the answer.