Visible-firearm detection

Cameras you already own, watching for a firearm nobody has noticed yet.

Analysis runs on a box inside your building. When a firearm becomes visible on camera, the people you designate get the frame, the location, and a confidence band — on a response ladder you configure, not one we impose.

This detects firearms that a camera can see. It cannot detect a concealed weapon, and detection is not prevention. It will produce false alerts; you will see your own measured rate before you decide how the system should respond to anything.

The part most of this category gets wrong

You choose the response, not us

Alerts can go straight to staff, through a reviewer, or through a monitoring agent who checks the camera first. Each one trades speed against false alerts, and the screen where you choose shows your own numbers beside each option.

Cancelling actually works

Every incident has one large cancel control. It reverses what can be reversed, asks every provider to stand down, sends an all-clear to everyone who was told, and records the whole round trip — including how far a dispatch got before you stopped it.

Video stays in your building

The appliance analyses streams locally and only ever uploads the cropped frame attached to a detection. There is no inbound connection to it, no cloud recording, and no face recognition anywhere in this product.

Cameras that cannot do this job are refused

A firearm sixty feet from a 1080p camera is about fifteen pixels tall. A low-resolution fisheye in a lobby is worse. During setup every camera is rated, and the ones that cannot resolve a firearm at their working distance are excluded from analysis and left off your invoice.

Billing for a camera that cannot work would mean charging for coverage you do not have, and quietly ruining the recall numbers for the cameras that do.