A single REST endpoint for invoice data extraction. Send a PDF or image, define the fields you want, and get back typed JSON — vendor name, invoice number, line items, totals, due dates. No templates, no training, no per-vendor setup.
Send the invoice and a schema — every field is returned with the type you declared. Strings stay strings, totals are numbers, dates are ISO-formatted.
curl -X POST \
https://api-parse.conversiontools.io/v1/extract \
-H "Authorization: Bearer YOUR_API_KEY" \
-F "file=@invoice.pdf" \
-F 'schema={
"vendor_name": "string",
"invoice_number": "string",
"invoice_date": "date",
"due_date": "date",
"line_items": [{
"description": "string",
"quantity": "number",
"unit_price": "number",
"amount": "number"
}],
"subtotal": "number",
"tax": "number",
"total": "number"
}'{
"status": "completed",
"pages": 1,
"data": {
"vendor_name": "Acme Corp",
"invoice_number": "INV-2026-0142",
"invoice_date": "2026-04-12",
"due_date": "2026-05-12",
"line_items": [
{
"description": "Consulting hours",
"quantity": 10,
"unit_price": 150.00,
"amount": 1500.00
}
],
"subtotal": 1500.00,
"tax": 150.00,
"total": 1650.00
}
}You define the fields once, the API guarantees the shape on every call. No null checks for "did the model return a string or an object this time?" — coercions are explicit.
Drop one sample invoice in the dashboard, get a full schema designed from your document. Edit, save, and reuse via the API. Skip writing JSON Schema by hand.
Works across every invoice format on the first call. Same schema, hundreds of vendors. The model reads layout and context — not pixel positions.
Generic OCR returns raw text. The Parse API returns structured JSON typed to your schema — invoice number, vendor name, line items as an array of objects, totals as numbers. You skip the post-processing step entirely.
No. The API uses a foundation model that already understands invoice structure across vendors and formats. You define the fields you care about in a schema and the API extracts them — no training data, no fine-tuning, no per-vendor templates.
Native PDFs, scanned PDFs, JPEG, PNG, WebP, TIFF. Multi-page invoices are processed end-to-end so line items that span pages get extracted correctly.
Single-page invoices typically return in under 5 seconds via the synchronous endpoint. Multi-page documents use the async polling endpoint and complete in under 30 seconds for most invoices.
Free tier: 100 pages per month with full API access. Starter at $29/mo gives 500 pages. Pro at $99/mo gives 5,000 pages with priority processing. No per-call fees, no overage charges.
Free tier covers your first integration. Upgrade only when you need more pages.