Work · 04 · Documents
0 copies
left behind after processing
Document anonymization without the cloud
DOCX · DOC · PDFwhat it accepts
1 containerinside the company: the anonymiser has no cloud, no AI services and no database
194automated tests, all passing
Availablestatus
The path of a document.
- 1 · youSend the file
- 2 · systemChecks the real format
- 3 · systemReplaces data with placeholders in Word, blacks it out in PDF
- 4 · systemRe-opens the result and scans it again
- 5 · systemSends the file back to the chat
- 6 · systemDrops the file and the replacement map
What is checked.
- ✓ No detected value stays readable
- ✓ Metadata is cleared, comment authors are replaced
What it refuses: the result cannot be guaranteed.
- Scans without text recognition
- Documents with tracked changes
- Hidden text
- Macros
- Password-protected PDFs
If the second scan fails.
- By default the file arrives with a warning
- A strict mode exists: then the file is not sent
A bot in the corporate messenger strips personal data from Word and PDF files entirely inside the company. The anonymiser itself has no cloud, no outside AI services and no database, and it drops the file and the replacement map after processing. The same line of bots includes document search in the e-document system, a list of pending approvals, email digests and meeting notes: one bot, one routine.
Area: Document workflow automation
Next piece of work: Changes to a BPM system without the chaos