Redaction that
stays local
Remove successfully matched names, SSNs, addresses, and account numbers from PDFs and plain text. Optional local OCR can process scans but may miss content. Runs entirely on your machine.
Optional local AI via Ollama can suggest context-dependent PII candidates such as client names, case captions, and custom identifiers. Models can miss or misclassify content, so deterministic rules and output review remain required. The model runs on your hardware.
Explore the PreviewOpen source. Not yet a production release.
Why Anonymize?
Rewrites matched text.
Matched PDF glyphs are removed and replacement text is written into a new file. Verify and inspect every result.
Stays on your computer.
Document processing stays local. We can't see your files because we never get them.
PDF & plain text.
Process text PDFs and plain-text exports. Optional local OCR can analyze scanned pages but may miss content.
How it works
Drop a file.
Drag a PDF or supported plain-text file into Anonymize.
Choose what to scrub.
Names, addresses, account numbers, and entity categories — choose exact values or load a saved profile.
Save the clean copy.
The original stays untouched. The new file rewrites successfully matched sensitive text; inspect it before sharing.
The engine is open-source under Apache-2.0. Audit the implementation and its documented limitations yourself.