Document Extraction API
One endpoint that turns invoices, receipts, and other business documents into structured JSON. Send a file, get named fields and line items back. Rule-based parsing reads every field, then an AI model checks the result and corrects what it finds wrong before you get it.
97%
97% field-level accuracy
Tested on 500+ real receipts and invoices
Every result also carries confidence and needs_review.
How we test →What it does
One API, many documents
Invoices and receipts through a single endpoint with a consistent response shape.
Structured, not raw text
Named fields and line items you can use directly — no post-processing of OCR text.
Confidence on every result
Each response includes a confidence score so you know what to trust and what to review.
Looking for a specific document type? See the Invoice Parser API and the Receipt OCR API.
Real output
Live example
Sample document
SellerNunez Inc
Invoice #31293737
Date2020-12-15
7 line items
Total$408.66
Real output from the API — nothing hand-edited. The data object returned by GET /api/v1/documents/{id}.
{
"document_type": "invoice",
"document_number": "31293737",
"invoice_id": "31293737",
"receipt_id": null,
"date": "2020-12-15",
"due_date": null,
"currency": "USD",
"merchant": "Nunez Inc",
"merchant_address": "6551 Janet Gardens, Michaelberg, KS 90952",
"merchant_email": null,
"customer": "Alvarez Inc",
"subtotal": "371.51",
"tax": "37.15",
"tax_rate": "10.00",
"discount": null,
"shipping": null,
"total": "408.66",
"payment_method": null,
"notes": null,
"line_items": [
{
"description": "Women's black cocktail dress, size 4",
"quantity": 3,
"price": "19.99",
"line_total": "59.97",
"vat_rate": "10.00"
},
{
"description": "Christmas Dress Grace Karin Sleeveless Christmas dress Red Black size Large",
"quantity": 1,
"price": "9.99",
"line_total": "9.99",
"vat_rate": "10.00"
},
{
"description": "Windsor womens black and Pink Bodycon Sleeveless Dress, Womens Size Small",
"quantity": 1,
"price": "15.83",
"line_total": "15.83",
"vat_rate": "10.00"
},
{
"description": "Lumiere Large Multicolored Orange Short Dress NWT",
"quantity": 4,
"price": "4.99",
"line_total": "19.96",
"vat_rate": "10.00"
},
{
"description": "Xhilaration Womens L Burgandy Boho Embroidered Festival Mini Dress NEW",
"quantity": 4,
"price": "14.95",
"line_total": "59.80",
"vat_rate": "10.00"
},
{
"description": "Johnny Was Medium Linen Wrap Dress Dessert Rose Embroiderd Bohemian",
"quantity": 2,
"price": "89.99",
"line_total": "179.98",
"vat_rate": "10.00"
},
{
"description": "Pre-Owned Women's Nah Nah Dress Size 8 Black Sheer Formal Sexy",
"quantity": 2,
"price": "12.99",
"line_total": "25.98",
"vat_rate": "10.00"
}
]
}Code
Submit the file, then fetch the JSON.
Every sample submits the file, then fetches the result until its status is completed or failed. On Starter and Pro, a webhook can deliver it instead.
# 1. Submit the document
curl -X POST https://docuparseapi.com/api/v1/extract \
-H "Authorization: Bearer $DOCUPARSE_API_KEY" \
-F "file=@invoice.pdf"
# → {"success": true, "document_id": "<id>", "status": "queued"}
# 2. Fetch the result (repeat until "status" is "completed" or "failed")
curl https://docuparseapi.com/api/v1/documents/<id> \
-H "Authorization: Bearer $DOCUPARSE_API_KEY"Try it on your own document
Upload a receipt or invoice and see the JSON. No sign-up needed.
Try it on your document — no sign-upOutput
Fields returned
document_typedocument_numberinvoice_idreceipt_iddatedue_datecurrencymerchantmerchant_addressmerchant_emailcustomersubtotaltaxtax_ratediscountshippingtotalpayment_methodnotesconfidenceneeds_reviewline_items[] (description, quantity, price, line_total, vat_rate)FAQ
Questions
Which documents can I send?
+
Any PDF, JPG, PNG or CSV, up to 10 pages per document. Receipts, invoices, purchase orders and expense reports get structured fields, AI correction and dashboard analytics.
What does document_type return?
+
invoice, receipt, purchase_order, expense_report or unknown.
Is it built for SaaS products?
+
Yes. One bearer token and one endpoint, the same JSON shape for every layout, and per-document billing that stays predictable as your users upload more.
How is AI used?
+
Rule-based parsing reads every field, then an AI model checks the result and corrects what it finds wrong before you get it.
How do high-volume teams use it?
+
Bulk uploads and batch processing up to 5,000 documents a month self-serve, with a spend dashboard to review results.
How long can a document be?
+
Every document includes up to 10 pages. A 10-page invoice counts as one document, the same as a one-page receipt. A document of 11 to 20 pages counts as two documents.
Are webhooks included?
+
On Starter and Pro. On Free, fetch results with GET /api/v1/documents/{id}.
How the extraction works+
- 01Read — the document is converted to text with OCR (for images) or parsed directly (for digital PDFs).
- 02Parse — deterministic rules pull out each field and line item and normalize dates and amounts.
- 03Check — the result is scored for confidence: totals are reconciled (subtotal + tax = total), required fields are verified, and formats are validated.
- 04Review — an AI model then reviews every document and corrects the fields it finds wrong, looking hardest at the ones the rules were unsure of.
Every response includes a confidence score and a needs_review flag, so you always know which results to trust and which to double-check.
Try it on your own document.
20 documents a month free, with full API access. No credit card. Webhooks on Starter and Pro.