Deepfake selfies and injected camera feeds defeat liveness checks that only look at pixels. The defence is not a better classifier. It is knowing that the frame arrived from a physical sensor at all.
Software that presents a video file to any app as though it were the camera is free and takes minutes to install. A liveness check downstream of it is examining the attacker's chosen frames.
Holding a phone up to a monitor defeats a surprising number of systems, and requires no technical skill at all.
Every honest applicant a liveness check turns away is a customer lost at the most expensive point in the funnel.
CapSeal checks the frame carries the metadata the operating system attaches to buffers that genuinely came off the sensor - camera intrinsics, sensor timestamps - which an injected surface has to fabricate consistently and usually does not.
Multi-frame parallax cross-checked against the device's own motion sensors. A screen and a printout are flat, and a flat subject cannot produce scene-consistent parallax.
Play Integrity and App Attest, bound into the manifest. That judgement comes from Google and Apple, not from us, and you can check it yourself.
CapSeal establishes that a selfie or document capture came from a physical camera pointed at a physical scene, on a device the platform vouches for.
The full position is in the Product Disclosure. We would rather you knew the boundary before you bought than after.
Try the verification half right now, with no signup, at capseal.ai/check.
Drop the capture SDK into your existing flow and keep your current liveness vendor.
Book a pilot Check a file first