Initializing Secure Environment…
Initializing Secure Environment…
Drawing a black rectangle over text hides it on screen and leaves it in the file. Anyone can select it, copy it, or run a text extractor, and that mistake has exposed court filings, government reports and corporate documents. This checker looks for every common way a redaction fails: text still under dark boxes, black annotation boxes stuck on top of the page, redaction marks that were never applied, and earlier versions of the file that still contain what you removed. It reads the file in your browser and never uploads it.
To check whether a PDF is properly redacted, test whether the text under each black box can still be extracted, whether redaction marks were applied, and whether an earlier saved version inside the file still holds the removed text. This checker does all three in your browser and lists any text it can recover, page by page. If it finds any, the document is not safely redacted.
The most common failure is cosmetic redaction: a black rectangle or highlight drawn over text in a word processor or PDF annotator. The page looks redacted and the text is still there. The second is the unapplied redaction, where someone marks areas for redaction and saves without applying them. The third is the incremental save, where the redacted version is appended to the file and the unredacted version stays inside it. The fourth is metadata: the title, author or comments can carry the very names a document was redacted to hide.
Proper redaction removes the content, not just covers it. Tools either delete the text and graphics within the area, or convert the page to an image with the area burned in black. Redact PDF on this site takes the second approach: pages with redaction boxes are rendered to images with the boxes filled, so no text exists underneath. That makes those pages unsearchable, which is the honest trade-off; run OCR afterwards if the rest of the page needs to be searchable.
It reads every page's drawing instructions to find dark filled boxes and compares their positions with the position of every piece of text. It checks annotations for redaction marks and black boxes, opens every earlier version stored in the file, and lists document properties. It does not read text inside images, and it skips text set at an angle. A clean result means no text was found under the boxes, not that the boxes cover the right things; read the document yourself before you send it.
Try to recover the hidden text. Select across the black box and paste it somewhere, search the document for a word you redacted, and check whether the file holds earlier versions. This checker automates all of that and lists anything recoverable. If it finds nothing under the boxes and no old versions, the text is gone from the text layer.
A PDF page is built from layers of instructions. A black rectangle is just another shape drawn on top; the text instructions underneath are untouched. A PDF viewer paints the box over the words, but copy, search and text extraction read the words directly. Real redaction deletes the text or replaces the page area with pixels.
Yes, if the file was saved incrementally. Some editors append changes to the end of the file instead of rewriting it, so the version before the redaction is still inside. The checker opens each earlier version and lists text that the final pages no longer contain.
Professional PDF editors redact in two steps: you mark areas, then you apply the redactions. A file saved between those steps has red or black outlines showing where redaction will happen, while the content underneath is still fully there. The checker lists every such mark.
No. It reads the text layer. On a scanned page with no text layer, black areas are part of the image itself, so there is nothing underneath to extract. If a scan has been OCR-processed, its text layer is checked like any other.
No. Everything runs in your browser. Documents that need redaction are exactly the ones that should not be uploaded to a third-party service.
More pdf tools — all free, no upload.