Skip to main content

Documentation Index

Fetch the complete documentation index at: https://vatextract.com/docs/llms.txt

Use this file to discover all available pages before exploring further.

VATextract uses advanced OCR technology to extract structured data from any invoice format—PDF, scanned documents, or images.

​
Supported File Formats

FormatDescription
PDFNative PDFs and scanned documents
JPEG/PNGHigh-resolution images
Multi-pageDocuments up to 100 pages

​
Extracted Fields

​
Header Information

FieldDescription
invoiceNumberUnique invoice identifier
invoiceDateDate the invoice was issued
dueDatePayment due date
deliveryDateGoods/services delivery date
purchaseOrderPurchase order reference

​
Financial Data

FieldDescription
netAmountPre-tax amount
vatAmountVAT/tax amount
vatRateVAT percentage
totalAmountTotal including tax
currencyISO currency code (EUR, GBP, USD, etc.)
freightAmountShipping/freight charges

​
Supplier Details

FieldDescription
supplierNameCompany name
supplierTaxIdVAT/Tax identification number
supplierAddressFull address
supplierContactEmail, phone, IBAN, etc.

​
Line Items

Each line item contains: 
{
  "description": "Product or service name",
  "quantity": 10,
  "unitPrice": 25.00,
  "amount": 250.00,
  "productCode": "SKU-12345",
  "vatCode": "T1"
}
VAT codes are automatically assigned based on the line item’s VAT rate and your accounting software. See VAT Codes for details.

​
OCR Providers

VATextract supports multiple OCR engines:
Default provider. Best for European invoices and complex layouts.
  • Excellent multi-language support
  • Strong table extraction
  • High accuracy on scanned documents
Configure your preferred OCR provider in Settings → Preferences, or set the OCR_PROVIDER environment variable for self-hosted deployments.

​
Extraction Confidence

Each extracted field includes a confidence score (0-100%). Low-confidence extractions are highlighted in the review interface for manual verification.

​
Geometry Data

For advanced integrations, VATextractprovides bounding box coordinates for each extracted field: 
{
  "fieldName": "TOTAL",
  "text": "$1,250.00",
  "confidence": 98.5,
  "geometry": {
    "boundingBox": {
      "left": 0.72,
      "top": 0.85,
      "width": 0.15,
      "height": 0.02
    },
    "pageNumber": 1
  }
}
This enables document overlay highlighting and programmatic field location.