_1772746035263-CVlnBaHI.png){kind=logo}
Live demo · Liquid Leak Detection
Try VisionAery on your own clip.
Upload an oilfield photo or short video clip and watch VisionAery’s edge liquid leak detection model outline pooled liquid in real time — the same analytic Twin Eagle runs at the well pad.
Connecting you to the demo…
About this Liquid Leak Detection demo
This hands-on demo runs the same VisionAery Liquid Leak Detection edge model that Twin Eagle Solutions deploys in the field. Upload a short clip from one of your own sites — a wellhead, tank battery, separator skid, pump, or containment area — and watch the analytic process it live. No account and no install are required, and you get a turn one at a time through a simple queue so the model has the full pipeline to itself during your run.
Liquid Leak Detection is a vision analytic, not a point sensor. It watches for the visual signatures of a release — pooling liquid, a spray, or a vapor plume — directly in the camera image, which means it can flag leaks across a whole scene rather than only at a single tapped fitting. In the demo you will see whether the edge model raised an alarm on your clip and, when it does, the supporting imagery the model captured at the moment of detection.
One thing the public demo intentionally leaves out is AVA, our automated verification step. A one-off clip from an unfamiliar site lacks the context AVA relies on — prior alarm history at that camera, a scene baseline, weather, and equipment metadata — so a verdict on a random video would not be trustworthy. In a production deployment AVA reviews every alarm against that site context before anyone is paged, which is how we keep nuisance alerts down without missing real events.
In the field the model runs on-premises at the edge, routes alerts through your existing video management system, and is tuned per site during a proof of concept. Field services are delivered across the lower 48, and VisionAery analytics are deployed more widely across the U.S., Canada, and internationally. To see how Liquid Leak Detection would perform on your own cameras, or .
About this Liquid Leak Detection demo
This page runs Twin Eagle's VisionAery Liquid Leak Detection model on footage you provide. Upload a short clip of an oilfield scene — a wellhead, tank battery, pump, valve, or flowline — and the same edge model we deploy in the field analyzes it and reports whether it raised a liquid-leak alarm. It is a hands-on look at how computer-vision leak detection actually behaves on real video, not a marketing animation or a synthetic simulation.
What Liquid Leak Detection does
VisionAery Liquid Leak Detection watches for the visual signatures of a liquid release — pooling, spray, sheen, and dripping — on standard and thermal cameras already at the site. In production it runs at the edge, on-device near the camera, so detection does not depend on streaming every frame to the cloud. The goal is to catch releases early, shorten response time, and reduce the environmental and cleanup cost of a leak that would otherwise run unseen between manual rounds.
How the demo works
When it is your turn, you upload one short clip and consent to processing it. The video streams to our pipeline, the edge model runs over it, and you see the live overlay followed by a result: either a leak alarm with the captured evidence, or an all-clear. To keep the public demo fair and responsive, it is single-session and queue-gated — you may see a short wait and a live preview while another visitor finishes.
Why verification is simplified here
In a real deployment, a detected alarm is cross-checked against per-site context — the camera's own history, a scene baseline, weather, and equipment metadata — before anyone is paged, which is what keeps false alarms low. A one-off clip from an unknown site has none of that context, so this public demo simply reports whether the edge model raised an alarm and skips the automated verification step that a live, tuned site would include.
From demo to a deployed site
A production rollout starts with a site survey and camera assessment, then field tuning on your own cameras during a proof of concept so you can judge accuracy before committing. From there the analytic is integrated with your alerting and SCADA so a confirmed leak reaches an operator the right way. Twin Eagle handles the full lifecycle — survey, install, integration, and ongoing support — across the lower 48, with VisionAery analytics deployed more broadly across the U.S., Canada, and internationally.