Document Scanning (OCR)

Extract structured fields from a photographed identity document without running a full verification. Use this when you only need the data on the document — for pre-filling forms or capturing reference details — and will handle authenticity, face match, and risk separately.

For full KYC with authenticity and face matching, use ID Document Verification instead.

🚀 Quick Start

curl -X POST https://api.afriroute.ai/api/v1/identity/scan-document \
  -H "Authorization: Bearer $AFRIROUTE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "document_type": "passport",
    "country": "NG",
    "image_front": "<base64-encoded-image>"
  }'

📡 Endpoint

Scan Document

POST /v1/identity/scan-document

Parameters

Field	Type	Required	Description
document_type	string	Yes	national_id, passport, drivers_license, or voter_id
country	string	Yes	ISO 3166-1 alpha-2 code
image_front	string	Yes	Base64-encoded front image
image_back	string	No	Base64-encoded back image (for IDs with rear data)
parse_mrz	boolean	No	Parse the Machine Readable Zone if present (default true)

💻 Code Samples

Node.js

const fs = require('fs');
const res = await fetch('https://api.afriroute.ai/api/v1/identity/scan-document', {
  method: 'POST',
  headers: { 'Authorization': 'Bearer $AFRIROUTE_API_KEY', 'Content-Type': 'application/json' },
  body: JSON.stringify({
    document_type: 'passport',
    country: 'NG',
    image_front: fs.readFileSync('passport.jpg', 'base64')
  })
});
const { fields } = await res.json();
console.log(fields.full_name, fields.document_number);

Python

import base64, requests

with open('passport.jpg', 'rb') as f:
    img = base64.b64encode(f.read()).decode()

res = requests.post(
    'https://api.afriroute.ai/api/v1/identity/scan-document',
    headers={'Authorization': 'Bearer $AFRIROUTE_API_KEY'},
    json={'document_type': 'passport', 'country': 'NG', 'image_front': img}
)
print(res.json()['fields'])

📊 Response

{
  "scan_id": "scan_8h2k9",
  "document_type": "passport",
  "country": "NG",
  "fields": {
    "full_name": "CHINEDU OKAFOR",
    "document_number": "A01234567",
    "date_of_birth": "1988-11-02",
    "sex": "M",
    "nationality": "NGA",
    "issue_date": "2021-03-01",
    "expiry_date": "2031-03-01"
  },
  "mrz": {
    "present": true,
    "valid": true,
    "raw": "P<NGAOKAFOR<<CHINEDU<<<<<<<<<<<<<<<<<<<<<<<<"
  },
  "field_confidence": {
    "full_name": 99.1,
    "document_number": 99.5,
    "date_of_birth": 98.7
  }
}

🗂️ Extracted Fields by Document

Document Type	Typical Fields
National ID	Name, ID number, DOB, sex, address
Passport	Name, passport number, nationality, DOB, expiry, MRZ
Driver's License	Name, license number, class, DOB, expiry
Voter ID	Name, voter number, DOB

💡 Best Practices

• Fill the frame with the document and avoid angled shots to maximize field confidence.
• Validate the MRZ — a mrz.valid: false strongly suggests a tampered or low-quality scan.
• Check field_confidence per field and prompt a re-capture when any critical field is below 90.
• Don't treat OCR as proof of authenticity — pair it with verification for KYC decisions.
• Strip the document image from your store after extraction unless retention is legally required.

⚠️ Error Handling

Code	HTTP	Description
IMAGE_QUALITY_LOW	422	Image too blurry, dark, or cropped to read
UNSUPPORTED_DOCUMENT	400	Document type not supported for that country
MRZ_UNREADABLE	422	MRZ present but could not be parsed
NO_DOCUMENT_DETECTED	422	No document found in the image

See the error code reference for the full list.

📚 Related Resources

• ID Document Verification
• Identity API Reference
• Risk Scoring

---

Last Updated: May 2026 | Need help? [email protected]