Automated CPF Extraction from Brazilian Driver Licenses

Why Extracting CPF from a CNH is Challenging

For businesses expanding into Brazil, capturing the CPF (Cadastro de Pessoas Físicas) is mandatory for tax, banking, and identity verification. However, extracting a CPF from a driver's license (CNH) isn't as simple as reading text. The CPF is printed in a dense block of numbers, often adjacent to the RG (State Identity Number) and the card's security registration numbers. Generic text-recognition tools frequently mix these strings up or fail to recognize the strict XXX.XXX.XXX-XX format when processing low-light or angled smartphone uploads. To understand how our purpose-built identity models solve this localization problem, explore our core driver license OCR infrastructure.

Precision CPF Parsing via API

Our API doesn't just read characters; it understands the topographical layout of the Brazilian CNH. By utilizing advanced semantic Key-Value mapping, it isolates the CPF with 99%+ accuracy, converting it into a clean, database-ready JSON format. The entire system operates on Cloudflare's global edge network, ensuring ultra-low latency for requests originating anywhere in South America. Better yet, integration is frictionless—you don't need to embed heavy SDKs into your frontend. You can test the accuracy immediately using the free image upload tool below.

High-Growth Use Cases for CPF Extraction

Integration & AI Prompts

Integrate CPF extraction in minutes. Our endpoint natively accepts direct file uploads, public URLs, or Base64-encoded strings (with built-in data cleansing for illegal whitespaces). For detailed payload structures, consult our developer documentation.

import requests
import base64

# 1. Prepare Base64 Image of the CNH
with open("brazilian_cnh_sample.jpg", "rb") as image_file:
    base64_image = base64.b64encode(image_file.read()).decode('utf-8')

url = "https://api.structocr.com/v1/driver-license"
headers = {
    "x-api-key": "YOUR_API_KEY",
    "Content-Type": "application/json"
}

# Direct Base64 injection
payload = {
    "img": base64_image 
}

response = requests.post(url, headers=headers, json=payload)
result = response.json()

if result.get('success'):
    data = result['data']
    print(f"✅ Extracted CPF: {data['personal_number']}")
    print(f"✅ Extracted RG: {data['extra_details']['rg_number']}")

{
  "success": true,
  "data": {
    "type": "drivers_license",
    "country_code": "BRA",
    "region": "SC",
    "document_number": "09876543210",
    "personal_number": "234.567.890-12",
    "surname": "PEREIRA",
    "given_names": "CARLOS EDUARDO",
    "sex": null,
    "date_of_birth": "1982-11-20",
    "place_of_birth": "FLORIANOPOLIS/SC",
    "date_of_issue": "2024-05-10",
    "date_of_expiry": "2029-05-09",
    "date_of_first_issue": "2000-08-25",
    "issuing_authority": "DETRAN SC",
    "address": null,
    "vehicle_class": "AB",
    "extra_details": {
      "fathers_name": "JOSE PEREIRA",
      "mothers_name": "MARIA FERREIRA",
      "rg_number": "3456789 SSP SC",
      "card_security_number": "0987654321"
    }
  }
}