# Paperwell Document Text & Data Scan

> Full text of a PDF, DOCX or scan plus identifiers, personal data, dates and amounts in it. Extracts the full plain text of a PDF (per page), DOCX, HTML, text file or photo/scan (OCR) and scans it for DOIs, ISBNs, ISSNs, ORCIDs, arXiv ids and PMIDs (with check-digit validation), personal data to redact (e-mail, IBAN, phone, payment card numbers), dates, amounts and periods. Paperwell uses it to make documents searchable, comparable and ready for redaction.

Part of usefulweb.site (https://usefulweb.site/). Results are deterministic calculations from the inputs and the listed assumptions; the API returns exactly what the web page shows.

## Use it
- Human interface: https://paperwelldocumentscan.usefulweb.site/ (interactive tool at https://paperwelldocumentscan.usefulweb.site/tool)
- JSON API: POST https://paperwelldocumentscan.usefulweb.site/api/v1/calculate with a JSON object of inputs (schema: https://paperwelldocumentscan.usefulweb.site/api/v1/tool)
- OpenAPI: https://paperwelldocumentscan.usefulweb.site/openapi.json
- MCP (Streamable HTTP): https://paperwelldocumentscan.usefulweb.site/mcp - tool `paperwell_document_scan`

## Inputs
- text (textarea): Text - Paste text, or upload a file below (exactly one).
- file (file): Or upload a file - PDF (text layer), DOCX, HTML, text or a photo/scan (PNG/JPG, read with OCR). Max 10 MiB. Read in memory and not stored.
- includeText (boolean): Include the full text in the result

## Method
- Text comes from the pasted text or the uploaded file: PDFs through their text layer (pdftotext, one form feed per page; scanned PDFs need OCR first), DOCX from its document XML, HTML without tags, images through Tesseract OCR (English).
- Identifiers are matched by pattern and validated by their check digits: ISBN-10 (mod 11), ISBN-13 (mod 10), ISSN (mod 11), ORCID (ISO 7064 MOD 11-2). DOIs, arXiv ids and PMIDs have no check digit and are reported as found.
- Personal data candidates: e-mail addresses, IBANs (only with a valid mod-97 checksum), payment card numbers (only with a valid Luhn checksum) and phone numbers (7–15 digits starting with + or 0). They become exact search terms and presets for the secure-pdf-redactor tool.
- Dates are recognised as 2026-03-12, 12.03.2026, 12 March 2026 / 12. März 2026 and March 12, 2026; slash dates like 03/04/2026 are reported without an ISO date because the order is ambiguous.

## Limits
- Pattern matching finds typical formats, not every personal detail: names, addresses and IDs in free text are not detected. Always review a document before sharing it.