Liveness

Liveness is passive presentation attack detection (PAD). You send a single face image and Sovera decides whether it is a real, present person rather than a photo, screen replay, or mask. It requires the verify:liveness scope.

How liveness works

Passive liveness needs no user action beyond looking at the camera. Sovera analyzes one frame and returns isLive plus a score. For a stronger anti-spoofing signal that also proves the user can respond to instructions, use active liveness, which adds an oval plus a turn-left or turn-right challenge.

Check liveness

Send a base64-encoded image. Pass a sessionId if this call belongs to a verification session.

POST

/v1/verify/liveness

IMAGE_B64=$(base64 -i selfie.jpg | tr -d '\n')

curl -X POST https://verify.lioncapventures.com/v1/verify/liveness \
  -H "Authorization: Bearer svk_live_xxxxxxxx" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: $(uuidgen)" \
  -d "{\"image\":\"$IMAGE_B64\"}"

Name
image
Type
string (required)
Description
Base64-encoded face image (JPEG or PNG).
Name
sessionId
Type
string
Description
Optional session this check belongs to, so the result is recorded on the session report.

Response

{
  "isLive": true,
  "score": 0.982,
  "threshold": 0.5,
  "latencyMs": 214,
  "detail": "passive_pad"
}

Name
isLive
Type
boolean
Description
Sovera's accept or reject decision for this frame using the provisional threshold.
Name
score
Type
number
Description
Confidence from 0 to 1 that the face is real and present.
Name
threshold
Type
number
Description
The cut-off score was compared against. Provisional pending in-house ROC calibration.
Name
latencyMs
Type
integer
Description
Server-side detection time in milliseconds.
Name
detail
Type
string
Description
A short label describing the detection path, useful for logging.

Thresholds are provisional. Keep the raw score and set your own accept and reject bands per use case rather than relying on isLive alone.