Biometrics

The production models, all of them

Six models, hardened inside a fraud-exposed identity business over eight years and trained on the same foundry the synthetic data comes from. Each ships as an on-prem container or SDK — no image leaves your infrastructure.

Book a call Licence terms

TessLabs Verify: 1:1 face verification

Annual licence · per environment

Confirms that two faces belong to the same person — a selfie against a document photo at onboarding, a login against an enrolment. Calibrated on MENAT-heavy evaluation sets, so thresholds hold on the populations most vendors underserve.

Returns: Match score, decision at your chosen threshold, quality flags.

Measured

Verification accuracy

97.53%

at a one-in-a-million false-match rate — FMR 1e-6, FNMR 2.4706%

On our own evaluation set.

TessLabs Identify: 1:N face identification

Annual licence · per environment

Searches one face against an enrolled gallery and returns ranked candidates — watchlist screening, duplicate-account detection, access control. Galleries to 10M identities; larger deployments scoped individually.

Returns: Ranked candidate list with scores, gallery-size-aware thresholds.

Measured

Identification accuracy

99.72%

at a 0.3% false-positive identification rate — FNIR 0.2763%

On our own evaluation set.

TessLabs Liveness: Passive presentation-attack detection

Annual licence · per environment

Scores whether the face in front of the camera is a live person or a presentation attack — print, screen replay, mask — from the capture itself, with no challenge gestures.

Returns: Liveness score 0.00–1.00, capture-quality flags.

Draft figures

Attack detection rate

99.9%

at the shipped operating threshold

Real faces wrongly rejected

0.1%

at the same threshold

Draft figures — to be replaced by the measured passive-liveness run.

TessLabs Detect: Deepfake & synthetic-face detection

Annual licence · per environment

Separates generated faces from camera-captured ones — deepfakes, face swaps and fully synthetic images. This is the discriminator side of the foundry, and the reason we can state that our own synthetics defeat frontier models while this engine still separates them.

Returns: Deepfake-likelihood score, generation-family signal, per-region heatmap.

Draft figures

Deepfake detection rate

99.9%

at the shipped operating threshold

Real faces wrongly flagged

0.1%

at the same threshold

Draft figures — to be replaced by the cleared detection run.

TessLabs Age: Age estimation

Annual licence · per environment

Estimates age from a single capture, reported against the product's own bins — 18–24, 25–34, 35–44, 45–59, 60+ — so a result maps onto the same bands the training data is built in.

Returns: Age bin with confidence, and a point estimate where one is required.

Measured

Age accuracy

98%

within ±3 years of actual age

On our own evaluation set.

TessLabs Gender: Gender classification

Annual licence · per environment

Classifies apparent gender from a single capture, evaluated per demographic group rather than on a pooled average, because a pooled number hides exactly the failures that matter.

Returns: Class with confidence, and the per-group breakdown behind it.

Measured

Gender accuracy

99%

overall accuracy

On our own evaluation set.

Evaluate before you licence

A 60-day evaluation licence, at no cost

Full API, capped gallery, non-production. Run it against your own traffic and your own thresholds before anything is signed.

Book a call