Verify Identity

POST/v1/verify

Upload a document image and a selfie for instant KYC verification. The API parses the document, compares faces between the document photo and selfie, runs passive liveness detection, and returns a verification result.

Requires: Starter plan or higher with kyc_verification feature enabled.

Request

Send as multipart/form-data.

Parameter	Type	Required	Description
`document`	file	Yes	Identity document image. JPEG, PNG, or WebP. Max 10 MB.
`selfie`	file	Yes	Selfie photo of the person. JPEG, PNG, or WebP. Max 10 MB.
`document_type`	string	No	Document category. Auto-detected if omitted.
`document_brand`	string	No	Regional variant. Auto-detected if omitted.
`scam_check`	boolean	No	Run scam database check on the extracted identity. Defaults to `false`.

Examples

curl -X POST https://api.dokai.dev/v1/verify \
  -H "Authorization: Bearer dk_live_your_api_key" \
  -F "document=@mykad.jpg" \
  -F "selfie=@selfie.jpg" \
  -F "document_type=identity_card" \
  -F "document_brand=mykad"

Response

Returns 200 OK with the verification result.

200 OK

{
  "id": "ver_abc123",
  "status": "verified",
  "type": "api",
  "document_type": "identity_card",
  "document_brand": "mykad",
  "face_match": {
    "matched": true,
    "score": 0.92,
    "threshold": 0.363
  },
  "liveness": {
    "passed": true,
    "score": 0.78,
    "checks": {
      "texture": 0.85,
      "frequency": 0.80,
      "reflection": 0.72,
      "sharpness": 0.92,
      "color": 0.75
    }
  },
  "identity": {
    "name": "AHMAD BIN IBRAHIM",
    "id_number": "901215-08-5523"
  },
  "metadata": {
    "processing_time_ms": 2340,
    "document_parse_id": "parse_xyz789"
  },
  "created_at": "2026-03-11T10:30:00Z",
  "expires_at": "2026-03-12T10:30:00Z"
}

Response fields

Field	Type	Description
`id`	string	Unique verification session identifier
`status`	string	`verified`, `unverified`, or `failed`
`face_match`	object	Face comparison result: `matched` (bool), `score` (0–1), `threshold`
`liveness`	object	Passive liveness result: `passed` (bool), `score` (0–1), `checks` (individual scores)
`identity`	object \| null	Extracted identity: `name`, `id_number`, `date_of_birth`, `nationality`
`metadata`	object	Processing details (time, model, linked document parse ID)
`created_at`	string	ISO 8601 timestamp
`expires_at`	string	When the result will be purged

Liveness checks

The liveness.checks object contains individual scores for each anti-spoofing heuristic:

Check	Description
`texture`	Skin texture analysis (LBP). Real skin has rich texture; printed photos are uniform.
`frequency`	Frequency domain analysis (FFT). Detects moiré patterns from printed or screen-displayed photos.
`reflection`	Specular reflection detection. Screens produce characteristic light reflections.
`sharpness`	Depth-of-field analysis (Laplacian). Natural photos have varying sharpness; reprints are uniformly blurred.
`color`	Color histogram analysis. Real skin has characteristic HSV distribution.

Retrieve a verification

GET/v1/verify/:id

Fetch a previously created verification result by its ID.

cURL

curl https://api.dokai.dev/v1/verify/ver_abc123 \
  -H "Authorization: Bearer dk_live_your_api_key"