Extract a payslip PDF to JSON
Everything below is the response the live API returned for a payslip, 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
Payroll integrations, income-verification providers and lenders who need proof of earnings.
The job, in the words people search for: “extract gross, deductions and net pay from a payslip”.
What makes this shape awkward
Earnings and deductions are two tables side by side, and side-by-side tables are the case ordinal column assignment is worst at — the two can merge into one six-column table if the gutter between them is narrow.
The measured result
| Pages read | 1 |
| Time | 10 ms (10 ms/page) |
| Labelled fields found unprompted | 5 |
| Tables | 2 (6 rows) |
| Input | payslip.pdf (28 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 payslip>",
"schema": {
"employee_number": "string",
"pay_date": "date",
"net_pay": "currency"
}
}'What came back
| Field | raw | value | confidence |
|---|---|---|---|
employee_number | EMP-04412 | "EMP-04412" | 60% |
pay_date | 2026-03-28 | "2026-03-28" | 70% |
net_pay | GBP 3,218.40 | 3218.4 GBP | 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.
Employee, Employee Number, Pay Period, Pay Date, Net Pay
Tables
Description · Units · Amount · 3 rows × 3 columns · page 1
Table confidence 75%. Header confidence 100% — the first row was promoted out of the data.
| Description | Units | Amount |
|---|---|---|
| Basic salary | 1 | 4,200.00 |
| Overtime | 6.5 | 214.50 |
| Shift allowance | 4 | 120.00 |
First 3 of 3 rows.
With options.rows_as_objects, the same row keyed by its header:
{
"Description": "Basic salary",
"Units": "1",
"Amount": "4,200.00"
}Description · Rate · Amount · 3 rows × 3 columns · page 1
Table confidence 75%. Header confidence 100% — the first row was promoted out of the data.
| Description | Rate | Amount |
|---|---|---|
| Income tax | 20% | 882.90 |
| National insurance | 8% | 353.16 |
| Pension | 5% | 220.50 |
First 3 of 3 rows.
With options.rows_as_objects, the same row keyed by its header:
{
"Description": "Income tax",
"Rate": "20%",
"Amount": "882.90"
}What this does not do
- No OCR. This reads the PDF’s text layer. A scanned or photographed payslip 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 payslip
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.