PDF Redaction & Verification

No. Each page is rendered to pixels, marked pixels are overwritten, and only the resulting images are embedded in a new PDF. No original page objects are copied. The output loses selectable text, accessibility tags, links, forms and layers.

The finished PDF is reopened and checked for extractable text, annotations and attachments. Rendered marked regions are checked for solid black pixels inside their boundaries. These checks do not establish that you found every secret, or that OCR cannot read content left outside a box. Review all pages visually.

Search proposes a box covering each matching PDF text run, not just a single word. It can match phrases across adjacent runs. This deliberately covers surrounding text to reduce missed glyphs. Scans, outlined text, unusual fonts or split reading order can prevent detection. Search is a helper, not a guarantee.

Save a plan, then reopen the same PDF and restore the plan. Plans contain coordinates and dimensions, not the original PDF or searched phrases. Review the restored marks before exporting.

Processing happens in your browser. Site advertising and optional analytics are described by the privacy policy. For highly sensitive material, assess whether an ad-supported website fits your needs.

32 MiB input, 50 pages, 500 marks, up to 12 million rendered pixels per page and 64 MiB of encoded page images per export. Choose 100–200 DPI to reduce memory. Cancel stops the export without a partial download.

No. The new image-only PDF is unsigned and untagged. It must not be presented as the original signed document. Remediate accessibility separately if required.