The Definitive Python SDK for Passport Data Extraction

Why Passport OCR is Difficult

Standard open-source OCR engines like Tesseract fail on real-world passport scans due to inherent complexities. Image defects such as camera glare, shadows, and non-uniform lighting degrade character recognition. Geometric distortions, including skew and rotation, misalign text regions, breaking parsing logic. The Machine Readable Zone (MRZ) itself, while standardized by ICAO 9303, requires precise parsing and checksum validation—a non-trivial task. Manually validating MRZ check digits or maintaining brittle RegEx patterns for dozens of international passport variations creates a significant engineering overhead that is both costly and error-prone.

Enterprise-Grade Extraction with StructOCR

StructOCR replaces fragile, multi-step OCR pipelines with a single API call. Our solution is built on pre-trained deep learning models specifically optimized for identity documents, not generic text. The API includes automatic image pre-processing, handling deskewing, denoising, and glare correction before analysis. Unlike Tesseract, which returns unstructured text blocks requiring post-processing, StructOCR directly outputs a standardized JSON object with validated fields. This eliminates the need for manual parsing, checksum calculations, and format normalization.

Production Use Cases

• Digital Onboarding (KYC): Reduce drop-off rates by pre-filling user data from Passports in < 2 seconds.
• Fraud Prevention: Detect tampered fonts or mismatched MRZ checksums automatically.
• Global Compliance: Handle Passports from 200+ jurisdictions without custom rules.

Implementation: Python SDK

The official Python SDK supports both MRZ and Visual Inspection Zone (VIZ) extraction. It automatically handles file encoding and parses fields like 'Place of Birth' that aren't available in the MRZ.

Prerequisite: pip install structocr

from structocr import StructOCR

# 💰 Save 30%+ vs competitors. Get 200 free requests instantly:
# 👉 https://structocr.com/register
# Initialize with your API Key
client = StructOCR("YOUR_API_KEY_HERE")

def scan_passport():
    # Note: Supports JPG, PNG, WebP (Max 4.5MB)
    image_path = "passport.jpg"

    try:
        print(f"Scanning {image_path}...")
        
        # The SDK handles file upload and API communication
        result = client.scan_passport(image_path)

        if result.get('success'):
            data = result['data']
            print("✅ Extraction Successful!")
            
            # Identity Data
            print(f"Passport #: {data.get('passport_number')}")
            print(f"Name:       {data.get('given_names')} {data.get('surname')}")
            print(f"Nation:     {data.get('nationality')} ({data.get('country_code')})")
            
            # Visual Zone Specifics (Not in MRZ)
            print(f"Birth Place:{data.get('place_of_birth')}")
            print(f"Issued At:  {data.get('place_of_issue')}")
            
            # Dates
            print(f"DOB:        {data.get('date_of_birth')} ({data.get('sex')})")
            print(f"Expiry:     {data.get('date_of_expiry')}")
        else:
            print(f"❌ Extraction Failed: {result.get('error')}")

    except Exception as e:
        # Handle SDK or Network errors
        print(f"An error occurred: {e}")

if __name__ == "__main__":
    scan_passport()

Sample JSON Output

StructOCR returns a normalized JSON object, regardless of the input image angle or quality.

{
  "success": true,
  "data": {
    "type": "passport",
    "country_code": "USA",
    "nationality": "UNITED STATES",
    "passport_number": "E12345678",
    "surname": "DOE",
    "given_names": "JOHN",
    "sex": "M",
    "date_of_birth": "1990-01-01",
    "place_of_birth": "NEW YORK, USA",
    "date_of_issue": "2020-01-01",
    "date_of_expiry": "2030-01-01",
    "place_of_issue": "PASSPORT AGENCY"
  }
}