Where removed text hides in a PDF
A PDF is a container, and the text you see on the page is only one place your words can live. When people say text “came back”, it was usually one of these:
| Hiding place | What this tool does |
|---|---|
| Under a black shape or highlight | Rebuilds the page from pixels, so the text is not in the new page at all |
| Earlier saved versions inside the same file | Writes a brand new file containing only the current pages |
| Title, author, subject, keywords | Removes document info and XMP metadata |
| Form field values | Flattens what prints onto the page, then drops the fields |
| Comments and sticky notes | Removes them |
| Bookmarks, alt text and tags | Removes them |
| Attachments | Removes them |
| Page thumbnails and app data | Removes them |
| Images shared between pages | Each kept page keeps only the images it actually draws |
How to remove text permanently
- Open the PDF above. It is processed on your device and not uploaded.
- Search for the text, or tick patterns such as SSNs or account numbers, then Find matches. Draw boxes over anything search cannot see, such as scanned text.
- Tap Redact and check.
- Read the check. If a string still turns up somewhere you did not mark, tap Redact every match and check again.
- Download when it shows 0 recoverable.
Why pages are rebuilt from pixels
Removing individual letters from inside a PDF’s drawing instructions is possible, but it is fragile: fonts, kerning, ligatures and odd encodings mean a tool can believe it removed a word while a fragment survives. Rendering the page and keeping only the image is blunt but certain. The trade-off is that text on a redacted page is no longer selectable; every other page stays as it was.
Want to see whether an existing PDF was redacted properly? Use the redaction checker.
Questions people ask
Is deleting text in a PDF editor permanent?
Often not. Editors that let you delete or retype text change what is drawn, but the old text can survive in an earlier revision of the file (PDFs are often saved by appending changes), in form field values, in bookmarks, in the document title, or in alt text. This tool builds a new file from scratch, so earlier revisions and leftovers are not carried over.
How do I know the text is really gone?
The tool checks the file it made before you download it: it extracts all text, decodes and searches every stream and string in the file, and confirms each redacted area is solid black. It shows the result as, for example, “Verified: 0 of 14 redacted strings recoverable”.
Can I remove text and leave white space instead of black?
Not in this version. The removed areas are black so a reader can see something was removed, which is the normal convention for redaction. White boxes also make it easy to miss that content was taken out.
Will the rest of the PDF change?
Pages without redactions are copied as they are. Hidden data is removed from the whole file: metadata, comments, links, form fields (their printed values stay on the page), attachments, scripts, bookmarks and page thumbnails.
What about text in the document properties?
The title, author, subject, keywords, creator, producer and dates are removed from every exported file, along with XMP metadata. These are a common leak: a file name or title often contains the very name that was redacted on the page.