DokMine — Use the document. Not the personal data.

Document PII · extract & redact<br>Use the document.<br>Not the personal data.

DokMine finds emails, phone numbers, URLs, IDs and names across PDFs,<br>Office files and images, then hands the file back redacted, in its<br>original format.

Start free<br>See how it works

Free tier · no card needed · files removed after processing.

Automated detection is never perfect — always review a redacted file<br>yourself before you share it.

contract_v2.pdfredacted

Blackout

Questions go to or by

phone at during business hours.

Typed tokens

Questions go to [EMAIL_1] or by

phone at [PHONE_1] during business hours.

Pull it out, or take it out

The same engine runs both ways. Extract to get a list of what it found<br>inside a document; redact to get the document back without it.

Extract<br>Extract emails, URLs and phone numbers

Point DokMine at a batch of PDFs, spreadsheets or mail archives and get<br>back the contact details and links it found inside — grouped by type and<br>listed per file. Handy for taking stock of what a document set contains,<br>auditing the links across it, or seeing what a file exposes before it<br>goes anywhere.

emails 42 found<br>urls 17 found<br>phones 08 found

Redact<br>Get the same file back, without the personal details

Choose how matches are masked: blacked out, or replaced with a<br>consistent typed token so the document still reads properly and the same<br>value maps to the same token throughout. Either way the file comes back<br>in its original format — a .docx as a .docx, a .pdf as a .pdf — with<br>layout and structure intact.

Contact [EMAIL_1] or call [PHONE_1].<br>Invoices go to [EMAIL_1] as well.<br>Portal: [URL_1]

Built for the parts everyone else skips

Spotting an email address is easy. Rewriting a .docx so the redaction<br>survives — including the text inside its embedded images — is not.

Format-preserving redaction

A redacted .docx comes back as a .docx, .xlsx as .xlsx, .pdf as .pdf. Layout, styling and structure stay intact — no flattening to plain text.

It looks inside images too

Screenshots and scans pasted into documents are easy to overlook. DokMine runs OCR over them and blacks out the matches in the image itself.

It tells you when it can't read something

If OCR can't run on an image, DokMine won't quietly hand the file back as though it had been checked — it flags that the image couldn't be read.

Checked, not just matched

Credit-card candidates are tested against a Luhn checksum and SSNs against structural rules, and common company names are filtered out of person-name results. Fewer false alarms to wade through.

Batch processing

Drop in a folder's worth of files at once. Results are itemised by file and entity type, so you can see what turned up where.

Short-lived by design

Uploads are processed to produce your result and nothing more. Download links expire automatically and files don't linger on the server.

What it looks for

Pick the entity types that matter for your document. Whatever you select<br>is searched for, listed for review, and redacted with a typed placeholder<br>— so [EMAIL_1] stays the same person throughout the file.

Consistent tokens per value, so redacted documents stay readable

Extraction-only mode when you just want to see what's inside

Review the hits yourself before you commit to a redacted copy

Email<br>Phone<br>URL<br>Names (people)<br>Credit card<br>SSN<br>IBAN<br>Passport no.<br>Date of birth<br>IP address<br>AWS key<br>API key / secret

Formats it reads

Office files, PDFs, mail, markup, code and scans — all through one pipeline.

Documents

PDF · DOCX · XLSX · PPTX — redacted in place, original format returned.

Mail & markup

EML · HTML · XML · JSON · YAML — body, headers and attachment text.

Text & code

TXT · CSV · MD · LOG · INI · PY · JS · CSS · Java · C/C++ — extension preserved.

Images & scans

PNG · JPG · TIFF · BMP · WEBP — read via OCR, then redacted visually.

Embedded media

Images inside Office documents are opened, OCR'd and redacted too.

Safety limits

Zip-bomb and oversized-extraction guards stop malformed uploads before they cost you a run.

Three steps, no setup

No install, no plugin, no configuration file. Open the workbench and drop a file in.

Drop your files

Drag one file or a whole batch onto the workbench. DokMine picks the right extractor per format automatically.

Choose what to look for

Toggle the entity types you care about, and pick extract-only or redact mode.

Review, then download

Check every hit, itemised by file and type, then download the redacted copy — and open it to confirm the result before you use it.

What DokMine can't promise

Worth reading before you rely on it. We'd rather be straight with you<br>than oversell what automated detection can do.

No guarantee of complete detection or removal

DokMine detects personal data automatically. No automated method is<br>exhaustive — it can miss things, and it can flag things that aren't<br>personal data at all.

Results vary with the quality, layout, language and encoding of the...