Verification Overview

Valydar provides a pipeline of checks that can be composed together. Each verification has a status that progresses through the pipeline.

Verification States

pending → completed | failed

State	Description
`pending`	Verification created, checks queued or in progress
`completed`	All checks finished
`failed`	One or more checks failed

Check	Endpoint	Description
Document	— (runs on upload)	OCR, MRZ extraction, format validation
Barcode	— (runs on upload)	Barcode detection, AAMVA parsing, MRZ cross-validation
Face Match	`/verifications/{id}/face-match`	Compare selfie to document photo
Reusable Face Auth	`/face/templates/*`	Enroll once, re-authenticate later with a selfie
Liveness	— (runs on selfie upload)	Passive selfie anti-spoofing (is it a real person?)
Active Liveness	`/verifications/{id}/active-liveness/*`	Challenge-response verification (blink, turn head, etc.)
Document Liveness	`/verifications/{id}/documents/{doc_id}/liveness`	Document authenticity (is it real or a copy?)
Document Forensics	— (runs on upload)	Tamper detection — splice/clone/recompression/typography
AML	— (runs in background)	Sanctions, PEP, and watchlist screening
KYB	— (runs in background)	Know Your Business — corporate entity verification
Affordability	— (runs in background)	Disposable income assessment for player protection

Each check returns a result with a boolean pass/fail:

{
  "passed": true,
  "score": 0.95,
  "details": {}
}

Field	Type	Description
`passed`	boolean	Whether the check passed
`score`	float	Confidence score (0.0 to 1.0, if applicable)
`details`	object	Check-specific data

Results are stored in checks_results keyed by check name (e.g., document, liveness, face_match).