Extract a commercial invoice PDF to JSON
Everything below is the response the live API returned for a commercial invoice, generated when this page was built. Not a description of what it would return — the output itself, including what it scored low on.
Who this is for
Accounts-payable tooling and B2B marketplaces, where invoices arrive as PDFs from hundreds of suppliers in nearly as many layouts.
The job, in the words people search for: “turn supplier invoices into rows in my accounting system”.
What makes this shape awkward
The header block is key/value pairs and the line items are a table, and the two need opposite treatment. A parser that sees everything as a table returns "Invoice Number:" as a one-column table and buries the line items under it.
The measured result
| Pages read | 1 |
| Time | 211 ms (211 ms/page) |
| Labelled fields found unprompted | 6 |
| Tables | 1 (9 rows) |
| Input | commercial-invoice.pdf (29 KB — synthetic, generated from a spec in the repo; real documents of this type cannot be published) |
Fields, with a schema
Ask for the fields you want by the label printed on the page. Each answer carries the text exactly as printed (raw), the value coerced to the type you asked for, a confidence, and a bounding box you can draw on the page to check it.
curl -X POST https://api.pdfcraft.dev/v1/extract \
-H "Authorization: Bearer $PDFCRAFT_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"file": "<base64 of your commercial invoice>",
"schema": {
"invoice_number": "string",
"invoice_date": "date",
"due_date": "date",
"total_due": "currency"
}
}'What came back
| Field | raw | value | confidence |
|---|---|---|---|
invoice_number | INV-2026-04817 | "INV-2026-04817" | 60% |
invoice_date | 2026-03-14 | "2026-03-14" | 70% |
due_date | 2026-04-13 | "2026-04-13" | 70% |
total_due | USD 14,882.50 | 14882.5 USD | 70% |
Every label it found without being asked
Send no schema and you get all of these, keyed by the label as printed. Useful for discovering what a new supplier’s layout actually contains before you write a schema against it.
Invoice Number, Invoice Date, Due Date, Purchase Order, Currency, Total Due
Tables
Description · Qty · Unit price · Line total · 9 rows × 4 columns · page 1
Table confidence 90%. Header confidence 100% — the first row was promoted out of the data.
| Description | Qty | Unit price | Line total |
|---|---|---|---|
| Steel bracket, 40mm | 6 | 288.00 | 1,728.00 |
| Aluminium housing | 9 | 107.00 | 963.00 |
| Cable loom, 2m | 18 | 801.00 | 14,418.00 |
First 3 of 9 rows.
With options.rows_as_objects, the same row keyed by its header:
{
"Description": "Steel bracket, 40mm",
"Qty": "6",
"Unit price": "288.00",
"Line total": "1,728.00"
}What this does not do
- No OCR. This reads the PDF’s text layer. A scanned or photographed commercial invoice has no text layer and returns
422 extraction_failedwith the page numbers that were blank — deliberately, and quickly, so you can route it somewhere that does OCR rather than waiting on a guess. - No model. Run the same file twice and you get the same bytes back. That is the trade: it cannot infer a field that is not printed, and it cannot hallucinate one either.
- Ambiguity is reported, not resolved. A date like 03/04/2026 stays a string, and a bare
$returns a null currency, because guessing wrong on either is not something you can recover from downstream.
Try it on your own commercial invoice
The playground takes a file and shows the same JSON, with every bounding box drawn over the page. No key and no signup for the first few; a free key gives you 100 pages a month.