Face Recognition · In development

Prove a real person is behind every signup.
Liveness, 1:1 matching, deepfake resistance.

In development: Passporter will capture a selfie, confirm the user is physically present with passive liveness, and match their face to the photo on their ID, blocking masks, replays and AI-generated deepfakes.

Passive
Liveness model (planned)
1:1
Selfie-to-ID match (planned)

Passive liveness detection

Planned: a single frame-stream confirms a live human is present without asking the user to blink, turn or smile.

  • No active gesture prompts
  • Presentation-attack detection

1:1 selfie-to-document match

Planned: the captured selfie is compared against the photo extracted from the user's ID and scored against a configurable threshold.

  • Configurable match threshold
  • Score returned per attempt
  • Document-photo extraction included

Deepfake & injection resistance

Planned: detection of AI-generated faces, virtual cameras and replayed video, so synthetic media cannot pass as a live capture.

  • Synthetic-face detection
  • Virtual-camera blocking
  • Replay & injection checks

Capture quality gating

Planned: the API scores framing, focus and lighting on every upload and rejects captures that cannot support a reliable match, so your own capture screen knows when to retry.

  • Quality score per upload
  • Retry guidance in the response
  • Works with any capture surface

Tunable security posture

Planned: raise the match threshold and liveness strictness for high-risk flows, or relax them for low-value onboarding, per integration.

  • Per-flow strictness
  • Risk-based step-up
  • Retry limits configurable

Privacy by design

Planned: biometric images processed for the match and discarded after the decision. Data is processed in the United Kingdom; another region is available by Enterprise contract.

  • Discard-after-decision option
  • Processed in the UK
  • GDPR-compliant handling

API Reference

Integrate in minutes.

Preview of the planned interface - this API is in development and not yet publicly available.

POST/api/v1/kyc/verifyPLANNED

Request

{
  "document_image": "<base64>",
  "selfie_image": "<base64>",
  "liveness_token": "lv_3a91f0",
  "match_threshold": 0.90
}

Response

{
  "verification_id": "kyc_9f3b2ce1",
  "decision": "APPROVE",
  "biometric": {
    "face_match_score": 0.97,
    "liveness": "PASSED",
    "deepfake": "NOT_DETECTED"
  }
}

Stop fraudsters at the selfie.

Face recognition is in development. Talk to us about early access and the onboarding flow you want to protect.