Primary line of work

AI defence work

Vigil is perception software. It watches live video, decides what is happening, grades how serious it is, and keeps the clip that proves it. It does not identify anybody, and that is deliberate.

See the pipeline FPV drone work →

Platform

01

Vigil

Motion detection fires on a bin bag in the wind. Vigil answers the question a person would ask: what is happening, how bad is it, and can you show me.

A running prototype today, with a local console, live camera input and a stored event history. It reads pose rather than pixels, which is what lets it tell a person who sat down from a person who went down.

01
Events, not motion
Vigil classifies what it is looking at — a person on the ground, someone who has stopped moving, a crowd breaking apart, an animal somewhere it should not be — instead of reporting that some pixels changed.
02
Severity, not alarms
Every event carries a grade. A person who sat down is not a person who collapsed, and the discriminator is stillness over time: somebody who chose to sit keeps shifting, and somebody who fell does not.
03
Many subjects at once
Every person in frame is tracked independently, and identity survives people crossing or passing behind an obstacle through colour re-identification. The count does not reset when two paths intersect.
04
Not only people
The same reasoning works on animals, because the features are self-referencing — proportions measured against the subject's own body rather than against a human template. Nothing in the geometry assumes two legs.
05
Evidence that holds
An event without its footage is an opinion. Unconfirmed events keep their clip for three days and then delete themselves; confirmed events are kept indefinitely. The record is the product, not a by-product.
06
Runs where the camera is
Inference runs on local hardware. There is no cloud round-trip in the detection path, and footage does not have to leave the site to be understood.

Vigil · from frame to record

Schematic
  1. 01 · InputFrameLive camera or file
  2. 02 · PoseKeypoints17 body points, per person
  3. 03 · TrackIdentityHeld across occlusion
  4. 04 · EventWhat happenedDown · still · dispersal
  5. 05 · GradeHow badSeverity, not a siren
Unconfirmed3 days, then gone
ConfirmedKept indefinitely

Prototype pipeline. Nothing in this chain identifies a person — see section 02.

Vigil · try it

Runs in your browser

Live demo

See it read a body

This runs the reasoning Vigil runs: find the body, decide what posture it is in, then watch whether it stops moving. Stand back so your whole body is in the frame, then work down the checklist.

Your camera opens inside this page. The video never leaves your device, nothing is recorded, nothing is uploaded, and no attempt is made to work out who you are.

Your browser will ask permission first. Stop the demo at any time and the camera is released immediately.

Not the Vigil model. Vigil runs YOLOv8-pose on local hardware; nothing of that kind runs in a browser tab, so this demo estimates pose with MediaPipe instead. The classification above it — posture, then stillness over time — is the same reasoning. A pose model is fetched from a CDN on first press, so the demo needs a connection.

Limits

02

What it will not do

Vigil does not identify anybody. No face recognition, no biometric matching, no watchlist. It classifies situations, not people.

This is a design decision, not a missing feature, and it is the one we get asked to reverse most often. Two reasons hold it in place.

01
Intent prediction is arithmetic, not tuning
Screening a crowd for who is about to do something runs into the base rate. At 99% specificity — far better than any fielded system — roughly 889 innocent people are flagged for every real one. No amount of model quality fixes a ratio set by how rare the event is.
02
Situations accuse nobody
A crowd suddenly dispersing is a stronger danger signal than any posture read off one individual, and it makes no claim about any person in it. That is the class of signal Vigil is built on.
03
Posture reporting stays off
The system can describe a body position, and that feature ships disabled. When it is enabled it reports a posture and a confidence — never a person, never an intention, and never the words "dangerous man".