Home / Data types / Documents
Real invoices and business documents for document AI
Datadhara sources real invoices, bills, purchase orders, catalogues and price lists from wholesalers, shops and suppliers: scanned or photographed, printed or handwritten. We transcribe and label them, remove vendor and customer identifiers, and license them for training document AI.
What we can source
| Record type | Format | Typical labels |
|---|---|---|
| Invoices and bills | Scans, phone photos | Line items, quantities, prices, totals, dates |
| Purchase orders | Scans, PDFs | Items, quantities, parties (masked) |
| Catalogues and price lists | Photos, PDFs | Product names, prices, units |
We collect to your spec, so volumes, languages and labels are agreed per project. Tell us what you need.
Example record
{
"invoice_no": "0417",
"date": "2026-10-03",
"vendor": "[REDACTED]",
"items": [{ "desc": "Rice 25kg", "qty": 4, "amount": 4600 }],
"total": 6440,
"consent_ref": "src-0192 / v2"
}
Values are illustrative. Real deliveries follow the schema agreed in your spec.
How personal information is removed
Names, phone numbers, addresses, tax IDs and bank details of vendors and customers are removed from the labels and masked on the images. See the full methodology.
What teams use it for
- OCR and key-value extraction
- Handwriting recognition
- Line-item table parsing
- Evaluation sets for document agents
How licensing works
Every source signs a written agreement that covers AI training. You start with a small pilot and a dataset card, and if the data works for you we agree volume, timeline and license terms (exclusive or non-exclusive) in writing.
Questions
Do you include handwritten documents?
Yes. Many small businesses still write bills by hand, and those are often the hardest cases for document models.
What labels come with the images?
Typically line items, quantities, prices, totals and dates, in JSON. Bounding boxes can be added if your spec needs them.
Tell us what your models are missing
A few lines is enough: the domain, the format, and roughly how much.
Request data →