Aller au contenu
Verify
Biometrics

Liveness that resists today’s spoof attacks

Active and passive liveness in a single capture — detects masks, screens, 3D prints, and AI-generated faces with iBeta Level 2 certification.

  • iBeta PAD Level 2 certified
  • Passive liveness with no gestures required
  • AI-generated face & deepfake detection
  • Sub-second SDK on commodity hardware
Get started
iBeta L2
PAD certified
<800ms
On-device latency
0.001%
Spoof acceptance
iBeta L2
PAD certified
<800ms
On-device latency
0.001%
Spoof acceptance
99.6%
Genuine accept rate
Capabilities

Built for production-grade trust

Each capability ships with monitoring, audit trails, and a typed SDK. Composable — pick what you need, leave the rest.

Passive liveness

Single-frame analysis with no gestures, prompts, or motion required.

Active challenge

Randomized blink, smile, or head-turn challenges for higher-assurance flows.

Anti-deepfake

Detect GAN/diffusion-generated faces and replayed video attacks.

1:1 face match

Match the live face to a document portrait or stored reference image.

How it works

From integration to verified result

A consistent flow whether you call via SDK, REST API, or trigger via dashboard.

  1. 1

    Capture

    User frames their face in the SDK overlay.

  2. 2

    Detect

    Liveness model runs on-device for sub-second response.

  3. 3

    Score

    PAD score, confidence, and detected attack type returned.

  4. 4

    Bind

    Optionally bind the live capture to a document portrait.

Use cases

Where teams deploy this

Patterns we see across regulated industries and high-velocity marketplaces.

Account re-verification

Step-up before high-risk actions like password reset or new device login.

Payouts & withdrawals

Confirm the right person is behind a large or unusual transaction.

Remote workforce

Authenticate contractors before they receive sensitive credentials.

Integrate in minutes

Production-ready SDKs and a typed REST API

Drop into your stack with a single dependency. Typed responses, idempotency keys, and structured errors out of the box.

liveness.tsxts
import { LivenessProbe } from '@verify/sdk-web';

<LivenessProbe
  mode="passive"
  onResult={(result) => {
    if (result.live && result.score > 0.9) {
      submit({ token: result.token });
    }
  }}
/>
FAQ

Frequently asked

Don't see your question? Reach out — we usually respond within a business day.

Ready to roll out facial liveness?

Spin up a sandbox in minutes, or talk to our team about a tailored rollout for your compliance and volume requirements.

Start free
Related solutions