hyperbots-inc/synthetic-stamped-documents

Synthetic Stamped Documents

100 business documents with 1 to 3 rubber stamps or seals each: PAID, RECEIVED, APPROVED, ENTERED, POSTED, FOR PAYMENT, COPY, DUPLICATE, CONFIDENTIAL, CANCELLED, VERIFIED and QC PASSED, plus circular company seals with ring text. The stamps come in red, blue, purple, green and black ink, rotated up to ±25°, at 55 to 90% opacity, with worn ink. Some have a handwritten date written inside. Stamps often land on top of totals, dates and line items, which is exactly where they cause extraction errors.

The base documents are invoices, tax invoices, purchase orders, delivery notes, bills of lading, credit notes, utility bills, receipts, remittance advice, NDAs, offer letters, work orders, packing lists and quotations.

Columns

Column	Description
image	1240 x 1754 JPEG
stamps	List of {type, shape, text, color, bbox, rotation_deg, overlaps_text, overlap_ratio, overlapped_blocks, handwritten_date}
fields	Printed fields {key, label, value, bbox} underneath the stamps
labels	JSON string of normalised gold values for the document
ocr_text	Printed text only (stamp text is in stamps[].text)

from datasets import load_dataset
ds = load_dataset("hyperbots-inc/synthetic-stamped-documents", split="test")
ex = ds[0]
print([(s["type"], s["color"], s["overlaps_text"]) for s in ex["stamps"]])

What to measure

Two things: whether extraction of the printed fields stays accurate under the stamps (compare with clean renders of the same documents), and whether the stamps themselves are detected and read correctly. A "PAID" or "RECEIVED 12 MAR 2025" stamp is often the evidence an AP workflow needs.

How the data was made

Every document is generated from a seeded random number generator by docgen.js and rendered on an HTML canvas in Chrome on macOS. Images use the system's print fonts and, for handwriting, the handwriting-style fonts installed with macOS (Bradley Hand, Noteworthy, Marker Felt, Chalkboard SE, Snell Roundhand, Apple Chancery, Savoye LET, Brush Script MT). Labels are taken from the generator itself, so there is no human annotation error. Field and block boxes are recorded at the moment each element is drawn.

All companies, people, addresses, account numbers and amounts are fictional. Phone numbers use ranges reserved for fiction where these exist (US 555-01xx, UK 07700 900xxx, AU 0491 570 1xx). Email addresses use the reserved example.com / example.org domains, and US ID numbers start with 9, which is never issued as an SSN.

To rebuild the dataset, load generator/docgen.js and generator/datasets.js in a browser page and run await DocSets.build('stamped'). Rendering depends on installed fonts, so pixels can differ on other systems, but text and labels are identical for the same seed.

Limitations

• The documents are synthetic. Their layouts, vocabulary and value ranges are much narrower than real enterprise traffic. Use them for benchmarking, regression tests and pipeline development, not as a replacement for evaluation on your own documents.
• All content is in English, with values formatted for 9 locales (US, UK, IN, CA, AU, DE, FR, SG, AE).
• Handwriting is rendered from fonts with per-character jitter. It is a stand-in for real handwriting, not a sample of it.

About HyperAPI

HyperAPI is Hyperbots' document intelligence API for enterprise documents. It provides parse, extract, classify, split and redact behind one endpoint and returns schema-validated JSON with a confidence score on every field. It has a REST API and a Python SDK, with an MCP server for agents available on request, and you can start free. See the HyperAPI Space for a walkthrough.

Part of the Hyperbots synthetic document suite

Dataset	Focus
synthetic-enterprise-doc-extraction	Field extraction across 31 enterprise document types
synthetic-doc-parsing	Layout blocks, reading order, Markdown and tables
synthetic-pii-redaction	PII spans and boxes for masking and de-identification
synthetic-doc-classification	31-class document classification
synthetic-doc-splitting	Multi-document bundles with true boundaries
synthetic-handwritten-forms	Forms filled in handwriting-style script
synthetic-stamped-documents (this dataset)	Stamps and seals over printed text
synthetic-bad-scans	Skew, blur, noise, shadows and compression at 3 severities
synthetic-dense-documents	Small type, multi-column layouts and long tables
synthetic-invoice-extraction	Invoice fields and line items (text + HTML)

License and citation

CC BY 4.0. Generator code is MIT.

@misc{hyperbots2026stampeddocs,
  title  = {Synthetic Stamped Documents},
  author = {Hyperbots Inc.},
  year   = {2026},
  url    = {https://huggingface.co/datasets/hyperbots-inc/synthetic-stamped-documents}
}