Liveness Detection API

0 reviews

Liveness Detection API adds presentation attack detection to face recognition systems and eKYC (online identity verification) flows. It determines whether the face in an image belongs to a live person or is a spoofing attempt: a printed photo, or a photo shown on a smartphone or tablet screen. Send a Base64-encoded JPEG or PNG image (up to 10 MB) to a single endpoint and receive a JSON response with a real-or-spoof verdict, a spoofing likelihood score from 0.0 to 1.0, face brightness, and the detected face position. There is no client library to embed and no per-OS binaries to manage, so it fits web applications and server-side identity verification flows. The service is operated with a target response time of 500 ms or less per request. The underlying technology is already in production in Japan, in authentication solutions for major medical institutions, access control terminals, and time and attendance terminals. The latest model focuses on stable decisions in unfamiliar capture environments and against new attack patterns; in an internal evaluation on approximately one million frames, it achieved 99.9% detection accuracy (internal evaluation, not a third-party certified value). Start with the Free plan (50 calls per month). Sign-up is self-service, with no sales call or contract negotiation required.