Get the whitepaper!
The Test Data Imperative: Towards SAP Autonomous Enterprise

Free · Test data only

Free synthetic sample documents for testing

Bank statements, pay stubs, invoices, lab reports and more, made for synthetic customers who don't exist. Use them to test OCR, document AI and redaction tools. Every page is watermarked as test data, and every file comes with the true values it contains.

Browse documentsGenerate them in bulk

One customer, every document

Meet Ava Okafor. She doesn't exist, and her paperwork agrees with itself.

The net pay on her pay stub is the payroll deposit on her bank statement, and her address is the same on her lab report. That's what lets you test a workflow end to end, not one document at a time.

SYN-0042 · Ava Okafor1200 Example Avenue, Springfield
Bank statement · Sep 2026Deposits 2 × $3,250.00
Pay stub · 15 Sep 2026Net pay $3,250.00
Lab report · blood panelSame address

The library

Each free sample: a watermarked PDF in clean, scanned and fax quality, and a JSON file of the true values. Other types are made in your own templates on request.

AllFinanceInsuranceHealth
SYNTHETIC TEST DATA
Bank statementFinance
Monthly statement with eight transactions and running balances.Free download · PDF and true-value JSON
SYNTHETIC TEST DATA
Pay stubFinance
Semi-monthly pay with tax and benefit deductions.Free download · PDF and true-value JSON
SYNTHETIC TEST DATA
W-2 wage statementFinance
Annual wage and tax statement for 2025.On request · in your own templates
SYNTHETIC TEST DATA
InvoiceFinance
Supplier invoice with line items, tax and payment terms.On request · in your own templates
SYNTHETIC TEST DATA
Purchase orderFinance
Order with ship-to address, line items and approver.On request · in your own templates
SYNTHETIC TEST DATA
Insurance claim formInsurance
Property claim with incident narrative and payout details.On request · in your own templates
SYNTHETIC TEST DATA
Lab reportHealth
Blood panel with reference ranges and flagged results.Free download · PDF and true-value JSON
SYNTHETIC TEST DATA
SOAP noteHealth
Clinical note in subjective, objective, assessment and plan form.On request · in your own templates
SYNTHETIC TEST DATA
Discharge summaryHealth
Hospital discharge with medicines and follow-up plan.On request · in your own templates

The library

Each free sample: a watermarked PDF in clean, scanned and fax quality, and a JSON file of the true values. Other types are made in your own templates on request.

AllFinanceInsuranceHealth
SYNTHETIC TEST DATA
Bank statementFinance
Monthly statement with eight transactions and running balances.Free download · PDF and true-value JSON
SYNTHETIC TEST DATA
Pay stubFinance
Semi-monthly pay with tax and benefit deductions.Free download · PDF and true-value JSON
SYNTHETIC TEST DATA
W-2 wage statementFinance
Annual wage and tax statement for 2025.On request · in your own templates
SYNTHETIC TEST DATA
InvoiceFinance
Supplier invoice with line items, tax and payment terms.On request · in your own templates
SYNTHETIC TEST DATA
Purchase orderFinance
Order with ship-to address, line items and approver.On request · in your own templates

The library

Each free sample: a watermarked PDF in clean, scanned and fax quality, and a JSON file of the true values. Other types are made in your own templates on request.

AllFinanceInsuranceHealth
SYNTHETIC TEST DATA
Insurance claim formInsurance
Property claim with incident narrative and payout details.On request · in your own templates

The library

Each free sample: a watermarked PDF in clean, scanned and fax quality, and a JSON file of the true values. Other types are made in your own templates on request.

AllFinanceInsuranceHealth
SYNTHETIC TEST DATA
Lab reportHealth
Blood panel with reference ranges and flagged results.Free download · PDF and true-value JSON
SYNTHETIC TEST DATA
SOAP noteHealth
Clinical note in subjective, objective, assessment and plan form.On request · in your own templates
SYNTHETIC TEST DATA
Discharge summaryHealth
Hospital discharge with medicines and follow-up plan.On request · in your own templates

How they're made

People first, then paperwork

  1. 1Synthetic customers. Names, addresses, employers and accounts are generated first, so every document about one person agrees.
  2. 2Templates filled. Synthesized's document generation fills each template with that customer's values.
  3. 3Truth recorded. A JSON file lists every value and the field it fills, so you can score an extraction tool field by field.

Rules every sample follows

  • Watermarked "Synthetic test data" on every page.
  • Fictional banks, employers and clinics. No real logos or signatures.
  • Social Security numbers from areas the SSA never issues (900–999).
  • UK phone numbers from Ofcom's drama range, 07700 900000 to 07700 900999.
  • Emails on example.com and example.net, domains reserved for examples.
  • For testing software only. Using them to deceive anyone is fraud.

Need 10,000 of these, in your own templates?

Synthetic document generation fills your templates with synthetic values, at the volume your test suite needs. The 48-hour proof shows it on your own templates first.

Get a 48-hour proofTesting document AI

Questions about the samples

Are these real documents?

No. The people, banks, employers and clinics are invented, and every page is watermarked. They look realistic enough to test software, and are marked so that no one could mistake them for real paperwork.

What's in the ground-truth file?

Every value on the document and the field it belongs to: account holder, period, balances, each transaction. Compare your tool's output with it to get accuracy per field.

Can I use them in my product or my own tests?

Yes. Every sample is released under CC BY 4.0: use it in your tests, your product or a benchmark, with credit to Synthesized.

Why do some numbers look odd?

On purpose. Account numbers end in 0000 and Social Security numbers start with 9, a range the SSA never issues, so no sample can match a real person's records.

The platform underneath, in production

Synthetic data from Synthesized has been tested against real data by the European Commission.

Documents and text use the same policies, the same engine and the same deployment as these results.

Banking · test data80%less time to provision test data, with compliant data generated through the API inside CI/CD.Deutsche BankCase studySynthetic data · fidelity0.359%modelled losses on Synthesized synthetic data with business rules, against 0.347% on the real data, in the European Commission JRC's evaluation.European Commission JRCReport

Published results from database and synthetic data work on the Synthesized platform.