The box you drew is just a shape sitting on top of the page. The text underneath is still there, untouched. Anyone can select that area, copy it, paste it somewhere and read exactly what you meant to hide. To actually hide it, the characters have to be taken out of the file, and that’s a different job from drawing over them.
Runs in this tab, on pdf.js, pdf-lib and qpdf (WebAssembly), so your file is read from your own device and never sent anywhere. Check it yourself
How do I fix it?
1Open the file in Redact PDFDrag a box over everything that has to go. Redaction deletes the characters inside each box and draws the black in their place.
2Check every page, not just the obvious oneIn my experience the same number turns up more than once: in a header, a reference line, the covering letter. The thumbnails make it quick to go through the whole document.
3Save, then try to copy from the resultOpen the saved file, select across the black area and paste. If nothing comes out, the words are really gone. That’s your proof.
4Clear the metadata before sendingRedaction deals with what’s on the page. Author names and editing dates live somewhere else in the file, and they go wherever the file goes.
Why does this happen?
Text and graphics are separate layers. A PDF keeps its words and its drawings apart. Painting over the words doesn’t change them at all, however solid the black looks on your screen.
Highlighting and marker tools are annotations. A black highlight is an annotation floating above the page. Whoever gets the file can click it, move it or delete it.
Exporting to PDF does not merge the layers. If you cover text in Word and then export, both things survive. The box is a shape, and the text is still text.
Screenshots hide the text but lose everything else. A picture of the page does get rid of the text, but it takes everything else with it. Nobody can search, select or copy any of the page after that.
Questions
Can someone really recover text under a black box?
Yes. They just select the area and paste it, or run any text extractor over the file. This has caused real leaks in court filings and company reports.
Does flattening the PDF make the box safe?
Flattening makes the box permanent, but the text underneath is still in the page and can still be pulled out. Redact first, then flatten if you want to.
What about a scanned page?
A scan has no text layer, so a box over it does cover the pixels. Flatten the file afterwards so the person you send it to can’t drag the box away.