What stops someone from registering with a genuine document that isn't theirs?

After the document check, the system matches the face of the person in front of the camera against the ID photo. If the faces don't match, identity isn't confirmed.

How does the check react to photos and deepfakes?

Passive liveness checks whether a live person is in front of the camera and catches attempts to pass the check with a photo, a mask or an on-screen image. The technology's resistance is certified to ISO 30107-3 Level 2.

Why check the face again before a payout?

Onboarding confirms who created the account. A repeat match before a payout or a change of payout details helps verify the same person is performing the sensitive action, even if the login and password ended up with an attacker.

Do you need to rerun the entire KYC?

No. To protect a payout it's enough to match the face against the template saved at onboarding. The marketplace gets a new biometric verdict without re-collecting the whole data set.

Does Biometric.Vision replace KYB, AML and transaction monitoring?

No. Biometric.Vision is responsible for document, liveness and face-match checks. Company verification, sanctions lists and operation monitoring stay in your compliance stack.

Can the check be embedded into an existing seller dashboard?

Yes. Connect via Web SDK, Mobile SDK, REST API or MCP — the check fits into your current seller flow without rebuilding your whole compliance stack.

Biometrics for e-commerce

One seller — one identity: from storefront onboarding to payout

Biometric.Vision links the document, the live face and the seller account into one thread. Someone else's document does not open the path to selling, and a hijacked login does not open the path to the money. An honest seller completes onboarding without long forms, and confirms before a payout or a change of payout details that the account is still run by them.

Start free Request a demo

Session · seller_onboardingSuccess

matchedface ↔ document

Document captureddata extracted

Liveness checkpassed

Face match with documentmatched

Ready for payoutstemplate saved

Seller lock · identity tied to the account until payout

The seller can go live on the storefrontThe verdict returns to your seller flow

Sound familiar?

A seller can't be kept at the door. But one document is no longer enough

Every extra step in registration pushes the first sale further away. But when the check is cut to a minimum, it opens a loophole for someone else's documents, stand-ins and deepfakes.

And onboarding is only the start. A real seller's account can be hijacked, payout details replaced, and a payout redirected. Buyers file complaints and chargebacks while the Trust & Safety team manually untangles the aftermath. That's why a marketplace needs a short path for the honest seller with continuous identity checks all the way to sensitive money actions.

Seller streamlive

Seller 4831Deepfake signs on camerarisk

Seller 4832Someone else's document · stand-in facerisk

Seller 4833Payout details changed · ATO riskreview

Seller 4834Document · liveness · facepass

Honest sellers and fraud share one stream — split them before the storefront and before the payout.

Document without a manual form

From document upload — straight to the seller check

The seller only needs to photograph a document. Biometric.Vision recognizes the data and checks the document itself — the team doesn't have to retype fields by hand or spot a forgery only after the first complaint.

433+ document types · 180+ countries — the marketplace gets structured data for the next step, and the seller reaches the storefront faster.

Session · seller documentCaptured

Document typerecognized

Document dataextracted

Portrait from documentready for match

Document in sessionData available to the next verification layer

A live person in front of the camera

A selfie has to be live, not just convincing

A photo, a mask or an on-screen deepfake can look convincing. Passive liveness detection catches an attempt to pass the check with this kind of substitution before the account gets seller status.

An honest user doesn't need to blink on command or turn their head: they just look at the camera. The technology's resistance to attacks is certified to ISO 30107-3 Level 2.

Step · passive livenessLayer passed

Photo instead of a facenot found

Recorded videonot found

Masknot found

Deepfake signsnot found

Video-stream substitutionnot found

Liveness confirmedISO 30107-3 Level 2 · no commands in front of the camera

Face and document

Someone else's passport won't open a store

Even a genuine document doesn't prove its holder is in front of the camera. Biometric.Vision matches the seller's selfie against the document photo. If the faces don't match, identity isn't confirmed and access to selling stays closed.

The algorithm ranks 37th in the world among 570+ solutions in the NIST FRVT ranking. Match accuracy is 99.9997%.

Step · face matchSuccess

Selfie

Document

A live person in front of the cameraconfirmed

Face match with documentmatched

Document holder confirmedAccess to selling can be granted

Identity before payout

Hijacking an account isn't enough to take the money

Before a payout or a change of payout details, the marketplace re-checks the face against the template saved at onboarding. A password and access to the dashboard alone are no longer enough to reach the funds.

The seller doesn't have to redo the full KYC. They confirm themselves by face, and the marketplace gets a verdict in 1.4 seconds.

Step · payout / payout detailsConfirmed

matchedface ↔ template

Match with onboardingmatched

Full KYC againnot needed

Verdict1.4 sec

The payout can go throughThe money goes to whoever completed onboarding

For developers

Classic integration

If your team prefers to assemble the architecture itself, connect ready SDKs and APIs.

JS · Webview

Web SDK / Flow Widget

A ready JS camera component for web registration or an in-app Webview. The check runs inside your onboarding; the design follows your brand book.

JavaScriptCustom designWebview

iOS · Android

Mobile SDK

A native verification flow in the mobile app: camera access and biometric micropacket transfer even on a weak connection.

SwiftKotlinOffline-ready

Backend

REST API

Layer results and the final verdict arrive in your backend logic: risk rules and the decision stay on your side. Auth via X-API-KEY.

RESTX-API-KEYServer-side

AI · AGENTS

MCP server for AI agents

Connect checks to agent scenarios without writing glue code. Works with Claude, GPT and any MCP client.

MCPClaudeAI-agents

seller_onboarding — REST API

# Seller onboarding / seller_onboarding

POST /v1/verification/check/

Host: api.biometric.vision

X-API-KEY: sk_live_••••••••••••••

Content-Type: application/json

{ "scenario": "seller_onboarding", "checks": ["document", "liveness", "face_match"] }

# → 200 OK

{ "status": "completed", "live": true, "face_match": "match" }

37 / 570+NIST FRVT rank

99.9997%match accuracy

ISO 30107-3Level 2 — liveness

ISO 27001No. 760281

Questions and answers

The scenario removes manual retyping of document data, and passive liveness removes camera commands. The seller goes through a sequential digital check without a long form.

500 checks a month free

Link storefront access and payout with one identity

Run through your own scenario: from onboarding to a repeat check before payout. See how the biometric lock fits into the seller dashboard without rebuilding your whole compliance stack.