DocArmor: AI-assisted PDF redaction with a human review step
What DocArmor processes
DocArmor is an AI-assisted redaction service for PDFs, scans and faxed documents. Its site describes OCR for image-only pages, visual detection for signatures and photographs, and barcode handling alongside text-based identification.
Users can describe material to remove in plain language, including a person’s role or relationship, rather than relying only on a fixed text pattern. That is useful when the sensitive information is contextual, but it also makes careful review of the proposed findings important.
How review and export work
The product’s workflow places findings in a checklist linked to their page locations. Reviewers can reject a finding, add a missing area, search for text and preview the redacted result before applying changes. DocArmor says its export process removes the selected content and checks every page.
Its site also describes stripping metadata, attachments and form values, along with an audit record of the template and review actions. These are vendor-described functions, not an independent certification that a particular uploaded document contains no remaining sensitive information.
Templates, deployment and a practical trial
The service lists templates associated with healthcare, education, court filings and public-records workflows. A template name does not by itself establish legal compliance for a specific disclosure. The intended recipient, document context and applicable organizational requirements still determine what needs review.
DocArmor advertises a seven-day, 20-page trial and several page-based plans, with separate enterprise deployment options. Confirm current retention settings and any required agreement before uploading confidential records.
For an initial trial, use a fictional document containing text, a scanned signature and a barcode. Compare the findings against a known answer list, inspect the export and test whether removed values remain extractable. That evaluates the complete review-and-sanitization workflow rather than judging only the appearance of black rectangles.