REST API · Developers

Invoice OCR Data Extraction API

Q: What invoice formats does the Invoice OCR API support?

PDF (scanned or digital), JPEG, PNG, TIFF, WebP images, multi-page documents. Structured formats: ZUGFeRD 2.x (hybrid PDF/XML), XRechnung 3.x, Peppol BIS 3.0, EDI 810, FatturaPA (Italy), Factura-e (Spain), GST e-invoice JSON (India, via IRP), CFDI 4.0 (Mexico). For structured XML/JSON formats we parse the embedded data directly — no OCR needed. For PDFs and images we use AI OCR with field-level confidence scoring.

Q: What is the accuracy and latency of the Invoice OCR API?

99.2% field-level extraction accuracy on standard invoice formats (tested across 50,000+ invoices from 40+ countries). Latency: under 2 seconds for single-page PDFs, under 5 seconds for multi-page invoices up to 20 pages. Structured format parsing (ZUGFeRD, Peppol, EDI 810) completes in under 500ms since no OCR is required.

POST a PDF or image. Get back vendor name, invoice number, line items, tax amounts, GSTIN/VAT/EIN, bank details — structured JSON, 99.2% accuracy, under 2 seconds. Any invoice format, any language.

REST APIPDF / ImageZUGFeRDPeppol BIS 3.0GST e-InvoiceEDI 810100+ Languages

Get API Key View API Spec

Request API Access

Sandbox with 500 free extractions. No credit card required for trial.

API Endpoints

Simple REST API

One endpoint for extraction. Multipart form-data with the invoice file. Response in JSON.

POST /api/v1/extract/invoice Extract all fields from one invoice

POST /api/v1/extract/invoice/batch Batch up to 100 invoices in one call

GET /api/v1/jobs/{job_id} Poll async batch job status

# POST /api/v1/extract/invoice
# Response JSON:
{
  "invoice_number": "INV-2026-00847",
  "invoice_date": "2026-08-15",
  "due_date": "2026-09-15",
  "vendor": {
    "name": "Acme Supplies Pvt Ltd",
    "gstin": "27AADCA1234B1Z5",
    "address": "Mumbai, Maharashtra"
  },
  "buyer": {
    "name": "XYZ Enterprises",
    "gstin": "07AABCX5678D1Z9"
  },
  "line_items": [
    { "description": "Industrial Motor 5HP", "hsn_code": "8501",
      "qty": 10, "unit_price": 8500.00,
      "igst_rate": 18, "igst_amount": 15300.00,
      "line_total": 100300.00 }
  ],
  "totals": { "subtotal": 85000.00, "tax_total": 15300.00, "grand_total": 100300.00 },
  "payment": { "bank_name": "HDFC Bank", "account_number": "XXXXXXX4521", "ifsc": "HDFC0001234" },
  "confidence": { "overall": 0.994, "invoice_number": 0.999, "gstin_vendor": 0.997 }
}

Extracted Fields

Every Field, Every Format

Field extraction adapts by country and format — GSTIN/HSN for India, VAT number/VIES for EU, EIN/TIN for USA.

🏢

Header Fields

Invoice number
Invoice date + due date
PO reference number
Currency + exchange rate
Payment terms

🏭

Party Details

Vendor name + address
GSTIN / VAT / EIN / ABN
Buyer name + GSTIN
Ship-to address
Bank account + IFSC / IBAN / BIC

📊

Line Items + Tax

Description + HSN/SAC code
Qty, unit, unit price
IGST/CGST/SGST or VAT
Discount + line total
Grand total + tax total

Pricing

API Pricing

Pay per extraction or fixed monthly — you own the deployment.

Starter

$0.08/doc

Setup fee: $5K–$12K

Up to 5,000 invoices/month
All standard formats
JSON output + confidence scores
REST API + webhook
500 free trial docs

Get API Key

Best Value

Monthly Fixed

$500/mo

Setup fee: $8K–$18K · Up to 10K docs/mo

10,000 invoices/month included
All formats incl. ZUGFeRD, Peppol
GL code suggestion module
Duplicate detection
ERP integration (SAP, Xero, QB)
SLA: 99.5% uptime

Start Monthly Plan →

Enterprise

Custom

100K+ invoices/month

Unlimited volume
On-premise / VPC deployment
Custom field training
Multi-language fine-tuning
SOC 2 / ISO 27001
Dedicated SLA

Request Enterprise Pricing

FAQ

Common Questions

What invoice formats does the Invoice OCR API support?

PDF (scanned or digital), JPEG, PNG, TIFF, WebP. Structured: ZUGFeRD 2.x, XRechnung 3.x, Peppol BIS 3.0, EDI 810, FatturaPA (Italy), GST e-invoice JSON (India), CFDI 4.0 (Mexico). For structured formats we parse embedded data directly — no OCR. For PDFs/images we use AI OCR with field-level confidence scoring.

What is the accuracy and latency?

99.2% field-level accuracy on standard invoice formats (tested on 50,000+ invoices from 40+ countries). Latency: under 2 seconds for single-page PDFs, under 5 seconds for multi-page invoices up to 20 pages. Structured XML/JSON parsing completes under 500ms.

Does it handle Indian GST invoices with GSTIN and HSN codes?

Yes — this is a core use case. The API extracts supplier GSTIN, buyer GSTIN, HSN/SAC codes, IGST/CGST/SGST amounts separately, IRN (Invoice Reference Number for e-invoice), and QR code data. It validates GSTIN format (15-character alphanumeric) and flags mismatches. GST e-invoice JSON from the IRP is also natively parsed.

Can I add custom fields to extract?

Yes. Enterprise and Monthly Fixed tiers support custom field definitions — you specify the field name, description, and example values. The model fine-tunes on your invoice layouts to find company-specific fields like internal cost center codes, project references, or approval stamps. Typical custom field training takes 3–5 business days.