OCR

SDK methods for identity document verification using optical character recognition and graded name matching.

The client.ocr namespace provides typed methods for verify identity (document OCR + profile name matching) and, where enabled for your tenant, processor configuration. The HTTP API is the source of truth for field names: documentUrl / documentBase64, profileName, expectedDocumentType.

SDK	Install	Namespace
TypeScript	npm install @caldera/platformxe-sdk	client.ocr
Python	pip install platformxe	client.ocr
Go	go get github.com/calderax/platformxe-go	client.Ocr

Verify identity document

Runs POST /api/v1/ocr/verify-identity: extracts text, matches profileName against OCR output, returns isVerified, nameReliabilityLevel, matchResult, requiresSecondaryValidation, verificationAdvisories, and extractedData.

See the full Verify Identity (OCR) reference for all fields, profileName.matchingContext (optional stateOfOriginCode, maritalStatus, etc.), and reliability tiers.

const { data } = await client.ocr.verifyIdentity({
  documentUrl: 'https://storage.example.com/uploads/nin-slip.jpg',
  profileName: {
    firstName: 'Ada',
    lastName: 'Okafor',
    middleName: 'Ngozi',
    idNumber: '98765432101',
    matchingContext: {
      stateOfOriginCode: 'DE',
      geoSignalSource: 'user',
    },
  },
  expectedDocumentType: 'NIN',
});

console.log(data.isVerified); // true only for FULL reliability + strong OCR
console.log(data.nameReliabilityLevel); // FULL | HIGH | MEDIUM | LOW | NONE
console.log(data.matchResult?.details?.regionalGeo); // when state code sent
console.log(data.requiresSecondaryValidation);

result = client.ocr.verify_identity(
    document_url="https://storage.example.com/uploads/nin-slip.jpg",
    profile_name={
        "firstName": "Ada",
        "lastName": "Okafor",
        "idNumber": "98765432101",
        "matchingContext": {"stateOfOriginCode": "AN"},
    },
    expected_document_type="NIN",
)

result, err := client.Ocr.VerifyIdentity(map[string]interface{}{
    "documentUrl": "https://storage.example.com/uploads/nin-slip.jpg",
    "profileName": map[string]interface{}{
        "firstName": "Ada",
        "lastName":  "Okafor",
        "matchingContext": map[string]interface{}{
            "stateOfOriginCode": "DE",
        },
    },
    "expectedDocumentType": "NIN",
})

Parameters (REST-aligned)

Field	Required	Description
documentUrl	One of URL/base64	Public HTTPS URL to the image
documentBase64	One of URL/base64	Base64-encoded image bytes
profileName	Yes	firstName, lastName, optional middleName, idNumber, matchingContext
expectedDocumentType	No	Document type hint (NIN, PASSPORT, …)

Python/Go parameter names may use snake_case — map to the same JSON body as the REST API.

Processor configuration

Where exposed for your organisation, processor endpoints tune OCR behaviour (confidence threshold, languages, supported types). See OCR processor.

Get processor config

const config = await client.ocr.getProcessor();

Update processor config

const updated = await client.ocr.updateProcessor({
  enabled: true,
  config: {
    confidenceThreshold: 0.9,
    supportedDocumentTypes: ['NIN', 'PASSPORT'],
  },
});

Auth

Use a valid PlatformXe API key (x-api-key) with the permissions required by your integration. Exact scope names depend on your key type — see Authentication.