Scanned PDFs: OCR on device, review before export
A scanned page is an image. Names and identifiers can exist only in pixels, not in a selectable text layer. AInonymizer checks supported scanned PDFs and document images with on-device OCR, proposes replacements and shows them for review.
Workflow
Open a supported scan, run local OCR-assisted detection, inspect highlights side by side, export and reopen the file. Check every page before sharing. A supported scanned PDF can receive a searchable redacted text layer when verification succeeds; otherwise it may remain image-only.
Limits
OCR may miss low-contrast, skewed, handwritten or very small text, stamps and dense tables. Face detection can miss faces. Complex layouts may need simpler text export or be blocked from export. AInonymizer does not guarantee anonymization; human review and your organization’s rules remain necessary. Currently in beta.