hyperbots-inc/synthetic-handwritten-forms

Synthetic Handwritten Forms

100 printed enterprise forms filled in by hand, using 8 handwriting-style scripts with per-character jitter, slant and baseline drift in blue or black ink. The forms are KYC and account opening, vendor bank details, credit applications, patient intake, insurance claims, expense reports, timesheets, maintenance reports and delivery notes. Ticked checkboxes, signatures and margin notes ("verified ID", "pay by Fri") are drawn by hand too.

HyperAPI runs handwritten and mixed print/handwritten content through the same vision-language parsing as print and gives every field a confidence score, so uncertain values can be sent for review. This set lets you measure that on forms where every value is known.

Columns

Column	Description
image	1240 x 1754 JPEG
writer_font, ink	Handwriting style used for the page
fields	List of {key, label, value, bbox, handwritten}: the gold value of every filled field and its box
checkboxes	List of {label, checked, options[{option, checked, bbox}]}
signatures	List of {signer, label, bbox}
annotations	Handwritten margin notes {text, bbox}
labels	JSON string of normalised gold values (ISO dates, "1,234.56" amounts)
ocr_text	Page text: printed labels plus handwritten field values (margin notes are in annotations)

from datasets import load_dataset
ds = load_dataset("hyperbots-inc/synthetic-handwritten-forms", split="test")
ex = ds[0]
print(ex["writer_font"], [(f["label"], f["value"]) for f in ex["fields"] if f["handwritten"]][:5])

Suggested metrics

Use field-level exact match and character error rate on handwritten values, checkbox state accuracy, and signature detection recall.

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('handwritten'). 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 (this dataset)	Forms filled in handwriting-style script
synthetic-stamped-documents	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{hyperbots2026handwrittenforms,
  title  = {Synthetic Handwritten Forms},
  author = {Hyperbots Inc.},
  year   = {2026},
  url    = {https://huggingface.co/datasets/hyperbots-inc/synthetic-handwritten-forms}
}