For the complete documentation index, see llms.txt. This page is also available as Markdown.

Liveness

Faceplugin Face Liveness Detection SDK and anti-spoofing. On-premise passive liveness for Android, iOS, Windows, and Linux Docker. POST /api/liveness on port 8084.

Overview

Faceplugin Face Liveness Detection SDK is an on-premise face anti-spoofing / presentation-attack detection (PAD) engine. It answers: is a live person in front of the camera, or a printout, phone screen, 3D mask, or deepfake-style video?

What it can do: Face SDK capabilities (Liveness section) — passive anti-spoofing, mobile vs server, and when to use it versus Face Recognition Identify.

This product is passive anti-spoofing: no smile / turn-head challenge. These docs do not claim iBeta certification. Contact Faceplugin if you need a certification statement. Active challenge demos on GitHub are not this product. Optional active prompts on some Face Recognition Identify flows also are not this SDK.

This SDK is standalone. It does not enroll people or run 1:N. If you already use Face Recognition Identify on mobile, that flow already includes passive 2D liveness. Choose this product when you need anti-spoofing without face matching.

Capable of detecting:

  • Printed photos

  • Screen replays

  • 3D models

  • Deepfakes (handled as one type of spoof attack — not a separate deepfake product)

Mobile apps run a live camera (VideoWorker plus faceDetection). Linux and Windows expose a Face Liveness API: one RGB JPEG over POST /api/liveness. Score 0.5 or higher is Real / pass.

Platforms

Start with Face SDK capabilities, then Mobile SDK or Server SDK.

FAQ

What is passive liveness detection? The engine scores a camera frame or JPEG without a user challenge animation. See Capabilities.

What is the difference between this and Face Recognition liveness? Face Recognition mobile Identify includes 2D liveness as part of matching. This product is anti-spoofing only.

Does it require the internet? No, after license activation.

Use cases

  • KYC / onboarding selfies — confirm a live person before face match (faceDetection / POST /api/liveness)

  • Banking and fintech — block photo and screen replay attacks on selfie capture

  • Access control — reject printouts and phone-screen spoofs at entry

  • Exam proctoring — score camera frames for presentation attacks without a challenge animation

  • Fraud prevention — JPEG anti-spoofing on your server (port 8084)

Last updated