Match Score
98%
Biometric Face Verification
With Liveness Built In
ID Analyzer's Biometric API matches the selfie against the ID photo, runs passive liveness, and flags deepfake or print-attack attempts — in a single call.
- Face match plus liveness in under 2 seconds.
- Passive liveness — no blinks required.
- Deepfake and presentation-attack detection.
Thwarting Attacks With Face Verification
Presentation attacks have evolved — high-res photos, screen replays, 3D masks, deepfake video. Texture, depth, and motion signals separate the live face in front of the camera from every spoofing variant trained against the model.
Blocks printed photos, screen replays, and silicone masks
Detects deepfakes and face-swap manipulation
Confidence-score output — set your own accept threshold
Four Biometric Checks, One Endpoint
Send the API two images — or one image and the previous face token — and receive a confidence score, liveness verdict, and presentation-attack signals in a single call.
1:1 Face Match
Submit two face images, get a confidence score between 0 and 1. Tune your accept threshold per flow — strict for high-value transactions, lenient for low-friction signups.
Passive Liveness
No blinks, no head turns. A single selfie is enough.
Active Liveness
Optional challenge-response — head turn or random prompt for high-risk flows.
Deepfake & Spoof Detection
Catches AI-generated face swaps, printed-photo replays, screen recaptures, and 3D mask attacks — in the same call as the match.
Four Surfaces, One Verification Engine
Same Biometric API, four very different jobs. Tune the match threshold and liveness mode per surface.
KYC Selfie-to-ID Match
Confirm the person uploading the document is the one in the photo. Strictest threshold, passive liveness on by default.
Workforce Sign-In
Replace the password prompt with a selfie. Active liveness on, low-latency mode.
Account Recovery
Match the recovery selfie against the enrollment photo. Stops social-engineering account takeover.
High-Value Transaction Step-Up
Trigger a face match before a wire, large transfer, or account-change action. One extra API call, fraud window closed.
From Selfie to Verdict in Four Steps
Every verification follows the same path. Sub-2-second response, structured JSON back, no human review in the loop.
1. Capture the Selfie
The user takes a live photo through their device camera, or you upload an existing face image. Any phone, laptop, or webcam.
2. Run Liveness
Texture, depth, and motion signals decide whether the face is in front of the camera or a photo, screen, or mask.
3. Match the Faces
1:1 comparison against the reference photo — the ID portrait, the enrollment selfie, or a stored face token.
4. Get the Verdict
JSON response — confidence score, liveness pass/fail, detected attack vectors. You decide whether to accept, review, or reject.
What Ships With Every API Call
Six capabilities baked into the verification engine. No tiered access, no separate billing for fraud signals.
Deepfake Detection
Models trained to spot AI-generated face swaps, GAN artifacts, and digital manipulation — the failure modes traditional face-match misses entirely.
Anti-Spoofing
Multiple detection layers for printed photos, screen replays, 3D masks, and video loops. Only live faces pass.
Cross-Platform
Any device with a camera — phone, tablet, laptop, kiosk. REST API plus SDKs for iOS, Android, Windows, macOS, and Linux.
High Accuracy, Low Bias
Neural networks trained against demographically diverse face data — accuracy holds across ethnicity and age groups.
Sub-Second Latency
Face match and liveness in under a second on the backend. The user never waits, the funnel stays warm.
Compliance Ready
ISO 27001 certified infrastructure, GDPR-aligned data handling, configurable retention, and regional processing options.
The Numbers That Matter to Procurement
Accuracy, latency, coverage. The metrics security and compliance teams ask about before they sign.
<2s
face match plus liveness, end to end.
190+
countries and territories covered by the paired ID engine.
ISO
27001:2022 certified infrastructure, awarded by BSI.
Questions From the Sales Call
Six questions that come up most often when teams evaluate the Biometric API. If yours isn't here, Talk to Sales in the menu reaches a human in under an hour.
Biometric face verification confirms a person's identity by comparing their live facial features against a reference photo — usually the portrait on an ID document. Unlike face recognition, which identifies an unknown face from a database, face verification is a 1:1 comparison: does this selfie match this reference?
Liveness analyzes the captured image for texture, depth, micro-motion, and reflection artifacts that distinguish a live face from a spoof attempt. The Biometric API ships with passive liveness — no blink prompts, no head-turn instructions — and an optional active-liveness mode for high-risk flows.
Yes. The engine includes dedicated deepfake-detection models that catch AI-generated and digitally manipulated faces. The models analyze facial artifacts, texture inconsistencies, and digital fingerprints characteristic of GAN-based generation.
Face spoofing is a presentation attack that tries to bypass verification by showing a fake version of an authorized person's face. Common forms: a printed photo, a face on a screen, a 3D silicone mask, a recorded video, or a deepfake. Liveness detection is the layer that counters all of these.
The Biometric API returns a confidence score per comparison. You set the accept threshold based on your risk tolerance — strict for high-value transactions, lenient for low-friction sign-ups. The matching model is trained on demographically diverse face data to keep bias low across ethnicity and age.
Yes. Use it on its own for face match and liveness, or combine it with the ID Verification API for end-to-end identity verification. The API accepts standard image formats (JPEG, PNG) or pre-extracted face tokens, and returns results as JSON.
Verify a Real Human Behind Every Signup
Free trial credits on signup, no card required. Your first face-match call lands in under an hour.