Docovue #

Privacy-first document scanning and OCR plugin for Flutter with smart auto-capture and anti-spoofing detection. Supports 12+ document types with 100% on-device processing – no cloud, no network calls, no data leaks.

✨ Features #

🎯 Smart Auto-Capture #

• Automatic document detection with edge recognition
• Live bounding box with color-coded feedback (White → Orange → Green)
• Hands-free capture – detects document and captures automatically
• Front & back camera support with seamless switching

🛡️ Anti-Spoofing / Liveness Detection #

• Prevents fraud – detects photos of photos, screen displays, photocopies
• Blur detection – ensures sharp, focused images
• Texture analysis – verifies physical document properties
• Reflection detection – identifies screen glare patterns

📄 Supported Documents #

🔒 Privacy & Compliance #

• ✅ 100% On-Device – All OCR runs locally (ML Kit on Android, Vision on iOS)
• ✅ Zero Network Calls – No data sent to cloud services
• ✅ No Data Storage – Plugin doesn't cache or persist data
• ✅ GDPR Ready – Built-in consent dialogs
• ✅ HIPAA Compatible – Suitable for healthcare apps (with proper implementation)
• ✅ PCI-DSS Friendly – Masks sensitive card data

🚀 Quick Start #

Installation #

Add to your pubspec.yaml:

dependencies:
  docovue: ^1.0.0

Run:

flutter pub get

Platform Setup #

Android

Add to android/app/build.gradle:

android {
    compileSdkVersion 34
    
    defaultConfig {
        minSdkVersion 21  // ML Kit requires API 21+
    }
}

dependencies {
    implementation 'com.google.mlkit:text-recognition:16.0.0'
}

Add to android/app/src/main/AndroidManifest.xml:

<uses-permission android:name="android.permission.CAMERA" />
<uses-feature android:name="android.hardware.camera" android:required="true" />

iOS

Add to ios/Runner/Info.plist:

<key>NSCameraUsageDescription</key>
<string>This app needs camera access to scan documents</string>
<key>NSPhotoLibraryUsageDescription</key>
<string>This app needs photo library access to import documents</string>

Minimum iOS version: 11.0

📖 Usage #

Basic Example – Auto-Capture Scanner #

import 'package:flutter/material.dart';
import 'package:docovue/docovue.dart';

class DocumentScannerPage extends StatelessWidget {
  @override
  Widget build(BuildContext context) {
    return Scaffold(
      appBar: AppBar(title: const Text('Scan Document')),
      body: DocovueScannerWidget(
        allowedTypes: const {
          DocovueDocumentType.aadhaar,
          DocovueDocumentType.pan,
          DocovueDocumentType.creditCard,
        },
        config: const DocovueScannerConfig(
          autoCapture: true,        // Enable smart auto-capture
          verifyOriginal: true,     // Enable anti-spoofing
          confidenceThreshold: 0.70,
        ),
        onResult: (result) {
          result.when(
            success: (document) {
              // Get top 3 extracted fields
              final summary = document.toSummary(maxFields: 3);
              print('Extracted: $summary');
              Navigator.pop(context, document);
            },
            unclear: (rawText, confidence) {
              print('Unclear (${(confidence * 100).toFixed(0)}%): $rawText');
            },
            error: (error) {
              print('Error: ${error.message}');
            },
          );
        },
      ),
    );
  }
}

Scan from File (Gallery/Existing Image) #

import 'package:image_picker/image_picker.dart';
import 'package:docovue/docovue.dart';

Future<void> scanFromGallery() async {
  // Pick image
  final picker = ImagePicker();
  final image = await picker.pickImage(source: ImageSource.gallery);
  
  if (image == null) return;

  // Scan document
  final result = await DocovueScanner.scanDocumentFromFile(
    imagePath: image.path,
    allowedTypes: const {
      DocovueDocumentType.aadhaar,
      DocovueDocumentType.passport,
    },
    config: const DocovueScannerConfig(
      verifyOriginal: false,  // Skip liveness for gallery images
      confidenceThreshold: 0.75,
    ),
  );

  result.when(
    success: (document) {
      final fields = document.toSummary(maxFields: 3);
      print('Success: $fields');
    },
    unclear: (rawText, confidence) {
      print('Could not classify document');
    },
    error: (error) {
      print('Scan error: ${error.message}');
    },
  );
}

Configuration Options #

// High-security mode (recommended for financial/ID documents)
DocovueScannerConfig.highSecurity()
// • verifyOriginal: true (anti-spoofing enabled)
// • confidenceThreshold: 0.90
// • maskSensitiveDataInLogs: true

// Healthcare mode (PHI/HIPAA)
DocovueScannerConfig.healthcare()
// • verifyOriginal: true
// • confidenceThreshold: 0.85

// Financial mode (PCI-DSS)
DocovueScannerConfig.financial()
// • verifyOriginal: true (fraud prevention)
// • confidenceThreshold: 0.95

// Development mode
DocovueScannerConfig.development()
// • verifyOriginal: false
// • debugMode: true

// Custom configuration
const DocovueScannerConfig(
  autoCapture: true,              // Smart auto-capture
  verifyOriginal: true,           // Anti-spoofing
  confidenceThreshold: 0.70,      // 70% minimum confidence
  showConsentDialog: true,        // GDPR consent
  maskSensitiveDataInLogs: true,  // Privacy protection
  debugMode: false,               // Debug logs
)

🎨 UI Components #

Auto-Capture Flow #

1. User opens scanner
   ├─ WHITE bounding box: "Position document"
   
2. Document detected
   ├─ ORANGE bounding box: "Hold steady... 1/4"
   ├─ Progress counter: 2/4 → 3/4 → 4/4
   
3. Document stable
   ├─ GREEN bounding box: "Capturing!"
   ├─ Auto-captures (2 seconds total)
   
4. Liveness check (if enabled)
   ├─ "Verifying authenticity..."
   ├─ ✅ Original → Continue to OCR
   ├─ ❌ Fake → "Original ID Required" error
   
5. OCR Processing
   ├─ Extracts text using ML Kit
   ├─ Classifies document type
   ├─ Extracts structured fields
   
6. Result
   ├─ SUCCESS: Returns top 3 fields
   ├─ UNCLEAR: Shows tips + raw OCR text
   └─ ERROR: Shows error message

Camera Controls #

• 🔄 Camera Switch: Toggle between front/back camera
• ⚪ AUTO Indicator: Shows progress (1/4, 2/4, 3/4, 4/4)
• 💡 Flash: Toggle flash/torch (back camera only)

🔧 Advanced Usage #

Extract Specific Document Fields #

result.when(
  success: (document) {
    document.map(
      aadhaar: (aadhaar) {
        print('Aadhaar: ${aadhaar.maskedAadhaarNumber}');
        print('Name: ${aadhaar.name}');
        print('DOB: ${aadhaar.dateOfBirth}');
      },
      pan: (pan) {
        print('PAN: ${pan.maskedPanNumber}');
        print('Name: ${pan.name}');
      },
      card: (card) {
        print('Card: ${card.maskedCardNumber}');
        print('Expiry: ${card.expiryDate}');
        print('Holder: ${card.cardHolderName}');
      },
      passport: (passport) => handlePassport(passport),
      generic: (generic) {
        // Generic extraction for other document types
        print('Fields: ${generic.detectedFields}');
      },
    );
  },
);

Validation Utilities #

import 'package:docovue/docovue.dart';

// Validate extracted data
bool isValidAadhaar = isValidAadhaar('123456789012');
bool isValidPAN = isValidPan('ABCDE1234F');
bool isValidCard = isValidCardNumber('4111111111111111');

// Mask sensitive data
String masked = maskCardNumber('4111111111111111'); 
// Output: "**** **** **** 1111"

String maskedAadhaar = maskAadhaarNumber('123456789012'); 
// Output: "XXXX-XXXX-9012"

🛡️ Security Best Practices #

Data Handling #

// ✅ DO: Use masked data in logs
print('Card scanned: ${document.maskedCardNumber}');

// ❌ DON'T: Log full sensitive data
print('Card: ${document.fullCardNumber}'); // Bad!

// ✅ DO: Enable data masking in production
DocovueScannerConfig(
  maskSensitiveDataInLogs: true,  // Always true in production
)

// ✅ DO: Enable anti-spoofing for ID/financial docs
DocovueScannerConfig(
  verifyOriginal: true,  // Prevents photo-of-photo attacks
)

Compliance #

GDPR – Show consent dialog:

DocovueScannerConfig(
  showConsentDialog: true,
  consentDialogTitle: 'Document Processing',
  consentDialogMessage: 'We process your document locally...',
)

HIPAA – Use healthcare mode:

DocovueScannerConfig.healthcare()

PCI-DSS – Never store full card numbers:

// ✅ Store masked version only
final safeData = card.toPciCompliantJson(); // Excludes full PAN

📊 Performance #

• OCR Speed: 1-3 seconds per document
• Auto-Capture: 2 seconds from detection to capture
• Memory: ~50-100 MB during processing
• Battery: Minimal impact (camera + ML Kit)

🐛 Troubleshooting #

"Unclear Document" Issues #

Causes:

• Poor lighting (too dark/bright)
• Document too far/close
• Blurry image (camera not focused)
• Low contrast document

Solutions:

// Lower confidence threshold
DocovueScannerConfig(
  confidenceThreshold: 0.60,  // Default: 0.70
)

// Enable debug mode to see raw OCR text
DocovueScannerConfig(
  debugMode: true,
)

Auto-Capture Not Triggering #

Check:

• Lighting is adequate
• Document fills 50-70% of bounding box
• Holding device steady for 2 seconds
• autoCapture: true in config

Liveness Detection Too Strict #

// Disable for testing
DocovueScannerConfig(
  verifyOriginal: false,
)

// Or use in specific scenarios only
verifyOriginal: documentType == DocovueDocumentType.creditCard,

🗺️ Roadmap #

• ❌ iOS implementation (currently Android-optimized)
• ❌ Enhanced MRZ parsing for passports
• ❌ Batch document scanning
• ❌ Custom document type training
• ❌ Improved low-light performance

📄 License #

MIT License - see LICENSE file for details.

🆘 Support #

Contact #

• Developer: Praveen
• Email: [email protected]
• Portfolio: praveen-dev.space
• Issues: GitHub Issues

Contributing #

Contributions welcome! Please:

1. Fork the repository
2. Create a feature branch
3. Commit your changes
4. Push to the branch
5. Open a Pull Request

⚠️ Disclaimer #

Docovue is not a legally certified KYC solution. OCR accuracy is best-effort. Always:

• Validate extracted data before use
• Implement proper error handling
• Follow compliance requirements for your jurisdiction
• Use human review for critical workflows

Made with ❤️ for privacy-conscious developers

Privacy-first document scanning and OCR plugin for Flutter. Supports Aadhaar, PAN, passports, credit cards, and more with 100% on-device processing.

🔒 Privacy & Compliance First #

Docovue is designed from the ground up for privacy and compliance:

• 🚫 No Network Calls: All OCR and processing happens on-device
• 🔐 No Data Storage: Plugin doesn't persist any document data
• 📱 No Analytics: Zero telemetry or tracking
• 🛡️ Compliance Ready: Built for PHI/HIPAA, GDPR, and PCI-DSS environments

✨ Features #

Supported Documents #

Legend:
⭐⭐⭐⭐⭐ Full Support: Dedicated extraction with high-confidence field-level validation
⭐⭐⭐⭐ Good Support: Returns as Generic with strong pattern matching
⭐⭐⭐ Generic Support: Returns top 3 detected fields from OCR text

Key Features #

• 🎯 High-Level API: Simple one-line document scanning
• 🔧 Low-Level API: Advanced OCR and classification control
• 📱 Live Camera Widget: Auto-capture with bounding box guides
• 🛡️ Anti-Spoofing: Liveness detection to verify physical documents
• 📊 Smart Output: Returns only top 3 most important fields
• ✅ Smart Validation: Luhn, Verhoeff, and format validation
• 🎨 Form Binding: Auto-populate form controllers
• 🌍 Multi-Language: Support for English, Hindi, and more
• ⚡ Fast Processing: Optimized on-device OCR

🚀 Quick Start #

Installation #

Add to your pubspec.yaml:

dependencies:
  docovue: ^1.0.0

Basic Usage #

import 'package:docovue/docovue.dart';

// Simple document scanning
final result = await DocovueScanner.scanDocument(
  context: context,
  allowedTypes: const {
    DocovueDocumentType.aadhaar,
    DocovueDocumentType.pan,
    DocovueDocumentType.creditCard,
  },
  config: const DocovueScannerConfig(
    showConsentDialog: true,
    maskSensitiveDataInLogs: true,
  ),
);

// Handle results
result.when(
  success: (document) {
    document.map(
      aadhaar: (aadhaar) {
        print('Aadhaar: ${aadhaar.maskedAadhaarNumber}');
        print('Name: ${aadhaar.name}');
        
        // Auto-populate form
        aadhaar.bindToForm(
          nameController: nameController,
          aadhaarController: aadhaarController,
          dobController: dobController,
        );
      },
      pan: (pan) => handlePanCard(pan),
      card: (card) => handleCreditCard(card),
      passport: (passport) => handlePassport(passport),
      generic: (generic) => handleGenericDocument(generic),
    );
  },
  unclear: (rawText, confidence) {
    // Show manual review UI
    showManualReview(rawText, confidence);
  },
  error: (error) {
    // Handle error
    showError(error.message);
  },
);

Using the Scanner Widget #

// Live camera scanner with auto-capture and liveness detection
DocovueScannerWidget(
  allowedTypes: const {
    DocovueDocumentType.aadhaar,
    DocovueDocumentType.pan,
    DocovueDocumentType.creditCard,
  },
  config: const DocovueScannerConfig(
    verifyOriginal: true,  // Enable anti-spoofing
    autoCapture: true,     // Enable auto-capture
    confidenceThreshold: 0.85,
  ),
  onResult: (result) {
    result.when(
      success: (document) {
        // Get only top 3 fields
        final summary = document.toSummary(maxFields: 3);
        print(summary); // {"number": "XXXX-1234", "name": "John", "dob": "01/01/1990"}
      },
      unclear: (rawText, confidence) => handleUnclear(),
      error: (error) => handleError(error),
    );
  },
)

Scan from File (Gallery/Existing Image) #

// Process an existing image file
final result = await DocovueScanner.scanDocumentFromFile(
  imagePath: '/path/to/image.jpg',
  allowedTypes: const {
    DocovueDocumentType.aadhaar,
    DocovueDocumentType.passport,
    DocovueDocumentType.invoice,
  },
  config: const DocovueScannerConfig(
    verifyOriginal: false, // Skip liveness check for gallery images
    confidenceThreshold: 0.75,
  ),
);

result.when(
  success: (document) {
    // Returns only top 3 most important fields
    final topFields = document.toSummary(maxFields: 3);
    print('Extracted: $topFields');
  },
  unclear: (rawText, confidence) {
    print('Could not classify document (${(confidence * 100).toFixed(0)}% confidence)');
  },
  error: (error) {
    print('Scan error: ${error.message}');
  },
);

📖 Usage #

Configuration Options #

// High-security configuration (recommended for production)
DocovueScannerConfig.highSecurity()
// ✅ verifyOriginal: true (liveness detection enabled)
// ✅ confidenceThreshold: 0.90
// ✅ maskSensitiveDataInLogs: true

// Healthcare/PHI environments
DocovueScannerConfig.healthcare()
// ✅ verifyOriginal: true
// ✅ confidenceThreshold: 0.85

// Financial/PCI environments  
DocovueScannerConfig.financial()
// ✅ verifyOriginal: true (anti-spoofing for card fraud)
// ✅ confidenceThreshold: 0.95

// Development configuration
DocovueScannerConfig.development()
// ⚠️ verifyOriginal: false (for testing)
// ⚠️ debugMode: true

// Custom configuration
const DocovueScannerConfig(
  showConsentDialog: true,
  maskSensitiveDataInLogs: true,
  debugMode: false,
  autoCapture: true,
  verifyOriginal: true,  // Enable liveness detection
  confidenceThreshold: 0.8,
  allowedLanguages: ['en', 'hi'],
  captureTimeout: Duration(seconds: 10),
)

Low-Level API #

For advanced use cases, you can use the low-level API:

// Extract text from image
final textBlocks = await DocovueScanner.extractTextFromImage(
  DocovueImageSource.camera,
);

// Classify document type
final documentType = await DocovueScanner.classifyDocument(textBlocks);

// Extract structured data
final document = await DocovueScanner.extractDocumentData(
  documentType,
  textBlocks,
);

Validation Utilities #

// Validate document numbers
bool isValidAadhaar = isValidAadhaar('123456789012');
bool isValidPAN = isValidPan('ABCDE1234F');
bool isValidCard = isValidCardNumber('4111111111111111');

// Mask sensitive data
String masked = maskCardNumber('4111111111111111'); // **** **** **** 1111
String maskedAadhaar = maskAadhaarNumber('123456789012'); // XXXX-XXXX-9012

🔒 Security & Privacy #

What Docovue Does #

✅ On-Device Processing: All OCR happens locally using ML Kit (Android) and Vision (iOS)
✅ No Network Calls: Zero HTTP requests or external API calls
✅ No Data Persistence: Plugin doesn't store images or extracted data
✅ Configurable Logging: Sensitive data masking in logs
✅ Consent Management: Built-in consent dialogs for GDPR compliance

What Docovue Does NOT Do #

❌ No Cloud OCR: No data sent to Google, AWS, or other cloud services
❌ No Analytics: No crash reporting, usage tracking, or telemetry
❌ No Background Processing: No data processing when app is backgrounded
❌ No Caching: No temporary storage of sensitive document data

Your Responsibilities #

As the app developer, you are responsible for:

• Data Handling: How you store, process, and transmit extracted data
• User Consent: Obtaining proper consent for document processing
• Compliance: Ensuring your app meets relevant regulations (GDPR, HIPAA, PCI-DSS)
• Security: Implementing proper encryption and access controls for sensitive data

🏥 PHI / HIPAA Considerations #

When processing healthcare documents:

// Use healthcare-optimized configuration
final config = DocovueScannerConfig.healthcare();

// Handle PHI data appropriately
result.when(
  success: (document) {
    // Ensure PHI is handled according to HIPAA requirements
    // - Encrypt data at rest and in transit
    // - Implement proper access controls
    // - Maintain audit logs
    // - Follow minimum necessary principle
  },
);

Important: Docovue provides the tools for secure document scanning, but HIPAA compliance requires proper implementation throughout your entire application and infrastructure.

🌍 GDPR Considerations #

For GDPR compliance:

const config = DocovueScannerConfig(
  showConsentDialog: true, // Required for GDPR
  consentDialogTitle: 'Document Processing Consent',
  consentDialogMessage: 'Your custom GDPR-compliant consent message...',
);

GDPR Rights: Your app must handle user rights including access, rectification, erasure, and data portability for any extracted document data you store.

💳 PCI-DSS & Card Scanning #

When scanning payment cards:

// Use financial-optimized configuration
final config = DocovueScannerConfig.financial();

result.when(
  success: (document) {
    document.map(
      card: (card) {
        // Use PCI-compliant data handling
        final safeData = card.toPciCompliantJson(); // Excludes full PAN
        
        // Never log full card numbers
        print('Card: ${card.maskedCardNumber}');
        
        // For PCI compliance:
        // - Never store full PAN unless you're PCI-compliant
        // - Use tokenization for card storage
        // - Implement proper network security
        // - Follow PCI-DSS requirements
      },
    );
  },
);

Warning: Card scanning is for UI convenience only. If you store or process card details, you must ensure your backend and application are PCI-DSS compliant.

🛠️ Platform Setup #

Android #

Add to android/app/build.gradle:

android {
    compileSdkVersion 34
    
    defaultConfig {
        minSdkVersion 21 // ML Kit requires API 21+
    }
}

dependencies {
    implementation 'com.google.mlkit:text-recognition:16.0.0'
}

Add to android/app/src/main/AndroidManifest.xml:

<uses-permission android:name="android.permission.CAMERA" />
<uses-feature android:name="android.hardware.camera" android:required="true" />

iOS #

Add to ios/Runner/Info.plist:

<key>NSCameraUsageDescription</key>
<string>This app uses the camera to scan documents for data extraction.</string>
<key>NSPhotoLibraryUsageDescription</key>
<string>This app accesses your photo library to scan documents for data extraction.</string>

Minimum iOS version: 11.0 (required for Vision framework)

🧪 Testing #

// Unit tests for validators
test('should validate Aadhaar numbers', () {
  expect(isValidAadhaar('123456789012'), isTrue);
  expect(isValidAadhaar('000000000000'), isFalse);
});

// Integration tests
testWidgets('should scan document successfully', (tester) async {
  await tester.pumpWidget(MyApp());
  
  // Simulate document scanning
  final result = await DocovueScanner.scanDocument(
    context: tester.element(find.byType(MaterialApp)),
    allowedTypes: {DocovueDocumentType.aadhaar},
  );
  
  expect(result.isSuccess, isTrue);
});

🚧 Limitations #

• Accuracy: OCR accuracy depends on image quality and document condition
• Languages: Limited language support (primarily English and Hindi)
• Document Variants: May not recognize all regional document variations
• Real-time: Not optimized for real-time video processing
• Offline Only: No cloud-based accuracy improvements

🗺️ Roadmap #

Version 1.1 #

• ❌ Enhanced MRZ parsing for passports
• ❌ Support for more Indian regional documents
• ❌ Improved address parsing
• ❌ Bank statement support

Version 1.2 #

• ❌ Real-time document detection
• ❌ Custom document type training
• ❌ Batch processing support
• ❌ Enhanced error recovery

Version 2.0 #

• ❌ Video-based scanning
• ❌ Document authenticity checks
• ❌ Advanced image preprocessing
• ❌ Custom ML model support

🤝 Contributing #

We welcome contributions! Please see our Contributing Guide for details.

Development Setup #

git clone https://github.com/prawinn04/docovue.git
cd docovue
flutter pub get
cd example
flutter run

📄 License #

This project is licensed under the MIT License - see the LICENSE file for details.

⚠️ Disclaimer #

Docovue is not a legally certified KYC solution. Extraction accuracy is best effort and human review is recommended for critical workflows. Always validate extracted data before using it in production systems.

🆘 Support #

• 📖 Documentation
• 🐛 Issue Tracker
• 💬 Discussions
• 📧 Email Support

Made with ❤️ for privacy-conscious developers