Mass surveillance in Connecticut

What officials told the public, against who actually searched. Step through the interactive map →

This site puts the record in one place: what each Connecticut police department searched, and what each town's officials have said about surveillance technology in their own public meetings. Every figure links back to its source — a released audit log, or the exact second of a meeting video — so you can check it yourself.

Want to do something about it? See flockoff.io.

Look at your town's police searches → Find your town in the matrix Questions? Spotted an error? Tell us

What's here

Your police department's searches 56 department dashboards Every licence-plate search each Connecticut department ran that reached six towns' cameras — filter by month, stated reason and how far it reached. Cross-town matrix 99 towns side by side Every town, every surveillance technology, mention counts from public meetings — each linking to that town's full report. This Week in Surveillance 7 days, refreshed What entered the record this week: meetings held, towns newly covered, and topics coming up for the first time in a town. Meetings to watch Next where it comes up Meetings already scheduled at bodies that keep returning to surveillance technology. For deciding which meeting is worth an evening. ALPR FOIA request template FOIA ask your own town A fill-in-the-blanks Connecticut FOI Act request covering Flock and the Axon in-car system most towns have never mentioned in public. Topic reports 17 technologies One page per technology — ALPR, drones, facial recognition, real-time crime centres — gathering what every town said about it. findings.csv CSV the whole dataset Every finding of mass surveillance technology in Connecticut's public meetings, in one machine-readable file: 7,004 rows, 101 towns, 18 topics, each linked to the exact second of the meeting video.

Explainers Soon coming soon Plain-language pages on how licence-plate readers actually work, one question per page. See below.

Read this before citing anything

The per-town findings are raw keyword matches, not verified claims. Keywords are ambiguous — "drone" can mean talking at length, "flock" a bird. Known false positives of that kind are removed, but a listed mention is still a pointer to a place in the record, not a finding in itself.

Absence of mentions is not evidence that a town has no surveillance technology. It often means the transcripts have not been fetched yet, that meetings were not recorded or captioned, or that the town's video archive lives on a platform this pipeline does not read. Every report states its own coverage — read the "Coverage status" section first.

Coming soon: explainers

Plain-language pages on how the data side of licence-plate readers works, one question per page. In preparation:

• Read the search record yourself — what a search event is, and three things about this data that are easy to get backwards
• What the camera actually records
• Hotlists, and what a "hit" means
• The search reason field
• Not the officer you know — who is actually searching your town's cameras
• Not the camera on the street
• Not lost puppies
• Not only Flock
• Nobody had to say yes to mass surveillance
• How this gets bought
• The transparency portal
• The four FOIA artifacts
• Reading a search log honestly
• Check your own town

Searching

Use the search box at the top of any page. It covers every town report, topic index and FOIA document on this site, and runs entirely in your browser.

Raw transcript archives

The plain-text meeting transcripts behind every report are published as release assets — one .zip per town, covering 92 Connecticut towns. Each archive contains a MANIFEST.txt giving its transcript count, the meeting dates it covers, and what share of that town's known meetings have actually been transcribed.

These are machine transcriptions of automatic captions, not official minutes. They contain mis-hearings, missing speaker attribution and dropped audio. Quote the video, not the transcript.

Two kinds of topic file

The distinction matters and the directories keep them apart:

• topics/<topic>.md — hand-curated cross-town indexes. These are claims, assembled deliberately, e.g. "cameras implemented with no legal process."
• topics/mentions/<topic>.md — generated per-technology mention reports. These are not claims: they are raw keyword matches with their source links attached, regenerated automatically.

Questions? Feedback?