The Gap in Most Document Tools

A text-only redaction tool redacts character runs. It cannot see a photograph on page four.

Documents are not text. A police report carries a booking photo. An insurance file carries a vehicle image with a plate in it. A medical record carries a scanned form with handwriting on it and a signature at the bottom.

Redactor rasterises document pages so the same detection that runs on images runs on the page content. A face, license plate or vehicle embedded inside a PDF is detected and obscured exactly as it would be in a standalone image.

How to Find What to Redact

Method	What it targets
Keyword	Specific words or phrases, wherever they appear
Pattern	Regular expressions and predefined templates such as SSN or credit card
AI-assisted	PII, text recognized through OCR, and objects such as faces and signatures

Templates for recurring types

Where documents share a layout, the regions carrying personal data sit in the same places. A template records those regions and the classes to detect, then applies to every document of that type, including across a bulk batch. A form that arrives every week is processed the same way each week rather than depending on which operator ran it.

What Document Redaction Covers

Text, including text with no text layer, recognized through OCR
Handwriting, through intelligent character recognition
Objects inside the document — faces, signatures, plates, vehicles in embedded images
Signatures, detected as their own class
Tables and spreadsheets, redacted by column and by row, so a whole field goes in one action rather than cell by cell
Non-Latin scripts — each text region is classified by script and routed to the engine that reads it: Latin (including Welsh), Perso-Arabic (Arabic, Urdu, Sindhi, Dari, Pashto), Devanagari, Cyrillic, Chinese and Japanese, Korean. Routing is per region, so a mixed-script page is handled region by region.

Compared With Adobe Acrobat

Acrobat	Redactor
One file at a time, one person at a time	Bulk batches, queued and worked overnight
Text only; an embedded photo is opaque to it	Page rasterised, so faces and plates inside PDFs are detected
Mark up each recurring form by hand, every time	Templates apply the pattern to every document of that type
No exemption codes, no coverage report	Exemption code on every redaction, printed on the output
Redaction correctness is the operator's memory	Custody trail with user, IP and timestamp

What Document Redaction Does Not Do

Redaction applies a solid fill on documents. Blur and pixelate are in development across the PDF, Office and CAD provider set.
Detection on embedded objects operates on the rendered page, so an object must be visible in the page as rendered.
Detection accuracy improves with clear, high-quality source documents.
A template assumes a stable layout. A document whose structure differs still needs review.
Script routing governs text recognition. PII language coverage is a separate and narrower matter.

How Document Redaction Is Evaluated

Three targeting methods: keyword, pattern including predefined SSN and card templates, and AI detection.
Formats span PDF, Word, Excel, PowerPoint, Visio, OneNote, CAD, email and HTML through three providers.
Detection runs on the rendered page, so faces, plates and vehicles inside embedded images are covered.
Spreadsheets redact by column or by row.
Text regions are classified by script and routed per region, so mixed-script documents redact.
Reusable templates make batch output consistent between operators.

Document Redaction Software