Receipt OCR API - Multi-Model Text Extraction

Extract structured data from receipt images using advanced OCR technology with support for multiple AI models including Google Vision, OpenAI, Azure, AWS Textract, Gemini, Hugging Face, DeepSeek, and Native OCR.

🌟 Features

Multi-Model OCR Support

Choose from 8 different OCR engines based on your needs:

• Google Vision API - High accuracy, excellent for printed receipts
• DeepSeek OCR - Advanced AI-powered text extraction
• Amazon Textract - Specialized for document and receipt analysis
• Azure AI Vision - Microsoft's computer vision service
• OpenAI GPT-4 Vision - State-of-the-art vision model
• Hugging Face - Open-source OCR models
• Google Gemini - Latest Google multimodal AI
• Native (Tesseract.js) - Free, no API key required

Intelligent Data Extraction

• Merchant Information: Name, address, contact details
• Transaction Details: Date, time, receipt number
• Financial Data: Total amount, subtotal, tax, currency
• Line Items: Individual items with prices
• Payment Method: Credit card, cash, etc.

Advanced Features

• ✅ Automatic Calculation Verification - Validates totals and tax amounts
• 📊 Batch Processing - Process multiple receipts simultaneously
• 🔄 Multi-Format Support - JPG, PNG, PDF files
• 📋 Structured JSON Output - Machine-readable data format
• 🎯 High Accuracy - Advanced parsing algorithms

🚀 Quick Start

Input Configuration

{
  "ocrModel": "native",
  "receiptUrls": [
    "https://example.com/receipt1.jpg",
    "https://example.com/receipt2.png"
  ],
  "extractLineItems": true,
  "verifyCalculations": true,
  "outputFormat": "detailed"
}

Required API Keys by Model

Model	Required Fields
Google Vision	googleVisionApiKey
DeepSeek OCR	deepseekApiKey
Amazon Textract	awsAccessKeyId, awsSecretAccessKey, awsRegion
Azure AI Vision	azureEndpoint, azureApiKey
OpenAI	openaiApiKey
Hugging Face	huggingfaceApiKey
Gemini	geminiApiKey
Native	None (uses Tesseract.js)

📋 Input Parameters

Required Parameters

• ocrModel (string) - OCR model to use

  • Options: google-vision, deepseek-ocr, amazon-textract, azure-vision, openai, huggingface, gemini, native
  • Default: native

• receiptUrls (array) - Array of receipt image URLs

  • Supports: HTTP/HTTPS URLs, data URLs, Apify key-value store URLs
  • Formats: JPG, PNG, PDF

Optional Parameters

• extractLineItems (boolean) - Extract individual line items

  • Default: true

• verifyCalculations (boolean) - Verify totals and tax calculations

  • Default: true

• outputFormat (string) - Output data format

  • Options: json (compact), detailed (with metadata)
  • Default: detailed

API Keys (Model-Specific)

Configure the appropriate API keys based on your selected OCR model. See the table above for required fields.

📤 Output Format

Detailed Output Example

{
  "receiptUrl": "https://example.com/receipt.jpg",
  "ocrModel": "google-vision",
  "success": true,
  "extractedAt": "2024-01-15T10:30:00.000Z",
  "merchantName": "SuperMart Store",
  "merchantAddress": "123 Main Street, City, State 12345",
  "date": "01/15/2024",
  "time": "10:25 AM",
  "receiptNumber": "TXN-12345",
  "currency": "USD",
  "subtotal": 45.50,
  "tax": 3.64,
  "totalAmount": 49.14,
  "paymentMethod": "Credit Card",
  "lineItems": [
    {
      "name": "Product A",
      "price": 15.99
    },
    {
      "name": "Product B",
      "price": 29.51
    }
  ],
  "calculationVerification": {
    "isValid": true,
    "errors": []
  },
  "metadata": {
    "confidence": 0.95,
    "processingTime": 1234,
    "imageSize": 524288
  },
  "rawText": "SuperMart Store\n123 Main Street..."
}

🎯 Use Cases

Expense Management

• Automate receipt data entry for expense reports
• Track business expenses in real-time
• Integrate with accounting software

Accounting & Bookkeeping

• Digitize paper receipts for record keeping
• Verify transaction details automatically
• Generate financial reports from receipt data

E-commerce & Retail

• Receipt verification systems
• Customer purchase tracking
• Warranty and return management

Fintech Applications

• Personal finance tracking apps
• Budget management tools
• Tax preparation software

🔧 Model Comparison

Model	Speed	Accuracy	Cost	Best For
Native	⚡⚡⚡	⭐⭐⭐	Free	Testing, low volume
Google Vision	⚡⚡	⭐⭐⭐⭐⭐	$$	High accuracy needs
Amazon Textract	⚡⚡	⭐⭐⭐⭐⭐	$$	Receipt-specific
OpenAI	⚡	⭐⭐⭐⭐⭐	$$$	Complex receipts
Azure Vision	⚡⚡	⭐⭐⭐⭐	$$	Microsoft ecosystem
Gemini	⚡⚡	⭐⭐⭐⭐	$$	Latest AI tech
DeepSeek	⚡⚡	⭐⭐⭐⭐	$$	Alternative to OpenAI
Hugging Face	⚡⚡	⭐⭐⭐	$	Open-source models

🔐 Security & Privacy

• All API keys are stored securely as secrets
• Images are processed in memory and not permanently stored
• Supports private/internal image URLs
• GDPR and data privacy compliant

💡 Tips for Best Results

1. Image Quality: Use high-resolution, well-lit images
2. Format: Straight, unfolded receipts work best
3. Contrast: Ensure good contrast between text and background
4. Model Selection:
   • Use Native for testing and low-volume processing
   • Use Google Vision or Textract for production workloads
   • Use OpenAI for complex or damaged receipts

🐛 Error Handling

The actor handles various error scenarios:

• Invalid or unreachable image URLs
• OCR processing failures
• Missing or invalid API keys
• Malformed receipt data

Each result includes a success field and error message when applicable.

📊 Batch Processing

Process multiple receipts in a single run:

{
  "ocrModel": "google-vision",
  "receiptUrls": [
    "https://example.com/receipt1.jpg",
    "https://example.com/receipt2.jpg",
    "https://example.com/receipt3.jpg"
  ]
}

The actor will process all receipts and provide individual results for each.

🔗 Integration

API Integration

const Apify = require('apify-client');

const client = new Apify.ApifyClient({
    token: 'YOUR_API_TOKEN',
});

const run = await client.actor('YOUR_ACTOR_ID').call({
    ocrModel: 'google-vision',
    googleVisionApiKey: 'YOUR_GOOGLE_API_KEY',
    receiptUrls: ['https://example.com/receipt.jpg'],
});

const { items } = await client.dataset(run.defaultDatasetId).listItems();
console.log(items);

Webhook Integration

Configure webhooks to receive results automatically when processing completes.

📈 Performance

• Processing Speed: 2-10 seconds per receipt (varies by model)
• Concurrent Processing: Up to 10 receipts simultaneously
• Maximum Image Size: 50MB per image
• Supported Formats: JPG, PNG, PDF

🆘 Support

For issues, questions, or feature requests:

• Check the Apify documentation
• Review the input schema for parameter details
• Ensure API keys are valid and have sufficient quota

🔄 Version History

v1.0.0

• Initial release
• Support for 8 OCR models
• Intelligent receipt parsing
• Batch processing
• Calculation verification
• Multi-format image support

---

Built with ❤️ using Apify Platform