Redact PDF — Truly Remove Sensitive Text

Yes — that is the entire point. The text beneath the boxes is deleted from the page content itself: the drawing commands that write it are removed from the file, so any normal text extraction (in-reader search, text extraction tools) finds nothing inside the covered area. The black cover afterwards is cosmetic — it shows something was removed there. And we verify it ourselves: after redaction we read the output file back and confirm no text remains inside any box, then report the result to you.

We refuse to apply it. If a box touches an image or a scan (where the text is pixels, not text), we decline the redaction and tell you which page and box caused it. Painting an image black without removing its pixels gives you a false sense of security — and that is exactly what we refuse to do. For that kind of content you need an image editor.

For text written inside the file: yes, with one sensible caveat. We remove the text so it can no longer be extracted, we strip the document metadata (title, author, XMP) and we review the output automatically. But we do not promise resistance to forensic analysis: anyone holding your original file or the embedded font could theoretically compare letter shapes, and any content that was never text in the first place (a photo of a document, say) is outside this tool's scope. Download the result and inspect the covered areas yourself before sharing it.

It is removed. After redaction we strip the document information dictionary (Title, Author, Subject and the rest), the XMP data and embedded files. Other interaction details such as form fields outside the boxes stay as they are — if you have a form holding sensitive data, place boxes over it too.

Because the safe direction is removing more, not less. Any line of text the box touches is removed in full, so no partial trace survives. That is why you should draw the box around everything you want gone — extra coverage is harmless, but leaving sensitive text behind is the real risk.

Yes — that is now the default mode. The selectable text on the page appears as shaded regions: click any region to mark it for redaction, click again to unmark it. Marked text is genuinely removed from the file by the same redaction engine. For non-text content (drawings or blank areas) switch to "Free box" and draw around it. Scanned pages show no selectable text because their content is an image, and true image redaction is not supported.

Yes. After you select text or draw a box on any page, an "Apply this page's areas to" bar appears — press "All pages" or set a range (from / to), check the preview of how many pages it covers, then press "Confirm application". The same areas are copied at the same relative position onto every page in the range (headers, footers and fixed fields are the usual case), and they all go through the same automatic post-redaction verification.

Yes. Type the word or phrase into the search box above the editor and hit search — we search every page and mark all occurrences automatically, ignoring hamza variants, diacritics and letter case in Arabic searches. The results are shown for review: click any occurrence to unmark it before running. If some occurrences cannot be measured precisely enough, they are left unmarked with a warning rather than redacted unreliably.

Yes. The "Preview redaction" button shows the current page before and after side by side, straight from the redaction engine itself: the translucent red area shows the full extent of what will be removed (selecting any part of a line removes the whole line), while black is the visible cover. The preview is approximate, not the final file — the actual output is verified automatically once the redaction runs.

No, not one byte. Previewing, redacting and verifying all happen inside your browser, and even the result file is built on your machine. No server upload, no logging, no daily limit.