Passive and active liveness

Before comparing faces, there's a different question: is a live person in front of the camera right now? We answer it with two layers that complement each other.

Try it with your face See the layers

Screen showing the next randomized gesture

Four layers, one decision

Passive looks at the image; active asks for gestures nobody knows in advance; the other two watch the entire video.

Passiveskin texture vs. paper and screens

Activerandomized gestures, in the right window

Continuitythe same face from start to finish

Integrityvirtual cameras and device swaps

Why randomize the gestures

session 5b0e…3 randomized gestures · 16.1 s

Smile5.2 s

Turn left6.1 s

Blink4.8 s

session 91a2…2 randomized gestures · 12.4 s

Open your mouth6.3 s

Raise your eyebrows6.1 s

Gestures, order and duration change every session. A video recorded in one doesn't fit the window of another.

In a virtual-camera injection test, a genuine video from an approved session, replayed into a new session, was rejected for exactly that reason.

Gestures per session2 to 3, randomized

Default catalog7 gestures

Window per gesture4.5 to 6.5 s

Countdown before recording2.5 s

What the person sees

Start screen with the script — **1 · The script**This session's gestures, randomized. Nothing records until they tap.

Get-ready countdown — **2 · Get ready**The first gesture shows up 2.5 s before recording starts.

**3 · Live guidance**The browser reads the face and says what's missing: "open your mouth wider".

All doneYou can close this page now.secure verification

4 · DoneThe page doesn't reveal the decision. It goes to your system.

Run your first verification and read the receipt.

Try it with your face Create account — 30 free per month