Will the deleted text still be in the file?

Not when the edit succeeds — the drawing instructions are removed, so the glyphs are gone. If a page cannot be edited precisely, the tool covers the text instead and warns you before export that the original wording is still extractable on that page.

Why does my edited text look different?

The original font is usually embedded as a subset containing only the characters the document already used. Replacement text needs glyphs that subset does not have, so it is drawn with a font this tool embeds. Shape and spacing will differ slightly.

Why doesn't the rest of the paragraph move?

A PDF positions every glyph individually and has no concept of a text flow. Nothing reflows when you edit, which is the same reason a PDF renders identically on every device.

Can I edit a scanned PDF?

Not the text, because a scan contains images rather than glyphs. You can cover or black out areas and use every page-level operation. There is no OCR step.

Does this work on password-protected PDFs?

No. Encrypted files are detected when you open them and rejected with an explanation, rather than being partially loaded.

Is this the same as redaction?

It removes the text, which is the substance of redaction, but a compliance workflow usually also demands an audit trail and metadata scrubbing that this tool does not provide. For regulated redaction, verify the output against your own requirements.

Text surgery

Edit PDF text

Delete a paragraph or rewrite its wording, with the original glyphs removed from the page rather than painted over.

How to delete or rewrite text in a PDF

1Open your PDFChoose the file from your device. Parsing happens locally, so the document is never transmitted.
2Pick a paragraphText blocks are detected from the page geometry and highlighted. Click the one you want to change.
3Delete or rewriteRemove the block outright, or type replacement wording in place of the original.
4Check the warningsIf a page could not be edited precisely, it is marked before export so you know that page was covered rather than cleared.
5DownloadExport rebuilds the PDF from the original bytes with your edits applied.

What works today

Right now this page gives you the overlay tools: draw a box to white out or black out any region, and place a text box anywhere to type replacement wording, including Chinese. Those work on every PDF, scans included, because they draw on top of the page rather than reaching into it.

Removing text from the file itself — so that it is gone rather than hidden — is still being built. The rest of this page explains why that is genuinely hard, what the difference means for you, and how to check the result once it lands. Until then, treat every box as a cover, not a deletion.

Why a PDF is not a Word document

Inside a PDF page there is no such thing as a paragraph. The page is a content stream: a list of drawing instructions that say things like "set this font at this size, move to these coordinates, draw these glyphs". Nothing records that a run of glyphs belongs to a sentence, a heading, or a column.

So a text editor for PDF has to reconstruct that structure. This one groups glyph runs into lines by their baseline, then groups lines into blocks when their spacing, left edge and font size are consistent. What you click on is an inference from geometry — usually a good one, occasionally off on complex layouts like tables or multi-column text.

Real deletion versus a white box

Most browser-based PDF editors implement "delete" by drawing a white rectangle over the words. The page looks right, but the text is still in the file: Ctrl+F finds it, selecting the area copies it, and a one-line command like pdftotext prints it back out. For anything sensitive, that is not deletion at all.

This tool removes the text-drawing instructions from the content stream instead, so the glyphs are genuinely gone. That requires matching each instruction to the text the renderer reports. When the match is confident, the text is deleted outright. When it is not — some documents nest their text inside form objects the parser does not follow — the tool falls back to a filled rectangle and says so, per page, before you export. It will not quietly hand you a file that only looks redacted.

How to verify it yourself

Do not take a tool's word for this, including this one. Open the exported file and search for a distinctive phrase you removed — if it is still findable, it was covered rather than deleted.

For a stricter check, run pdftotext edited.pdf - in a terminal and read the output. That bypasses the viewer entirely and prints whatever text the file still contains.

Why the font changes when you rewrite text

PDFs usually embed fonts as subsets: only the glyphs the document actually uses are included, to keep the file small. If the original never used the letter your replacement needs, that glyph simply is not in the file.

New text therefore has to be drawn with a font this tool embeds itself. The result stays legible and selectable, but letterforms, spacing and weight will not match the original exactly. Short edits usually blend in; replacing a whole heading tends to be noticeable.

Why the following text does not reflow

Every glyph in a PDF carries its own position. There is no paragraph object that owns a block of text and can re-lay it out, which is exactly why a PDF looks identical everywhere it is opened.

The consequence is that shortening a sentence leaves a gap and lengthening one runs into whatever sits below. Keep replacement text close to the original length, or delete and re-add the surrounding lines to suit.

Scanned PDFs have no text layer

A scan is a photograph wrapped in a PDF container. There are no glyphs to find and no instructions to remove, so text editing has nothing to operate on. Pages like this are labelled as scans and their text tools are disabled rather than silently doing nothing.

You can still cover or black out regions of a scan, and all the page-level operations work normally. This tool does not run OCR.

FAQ

Will the deleted text still be in the file?: Not when the edit succeeds — the drawing instructions are removed, so the glyphs are gone. If a page cannot be edited precisely, the tool covers the text instead and warns you before export that the original wording is still extractable on that page.
Why does my edited text look different?: The original font is usually embedded as a subset containing only the characters the document already used. Replacement text needs glyphs that subset does not have, so it is drawn with a font this tool embeds. Shape and spacing will differ slightly.
Why doesn't the rest of the paragraph move?: A PDF positions every glyph individually and has no concept of a text flow. Nothing reflows when you edit, which is the same reason a PDF renders identically on every device.
Can I edit a scanned PDF?: Not the text, because a scan contains images rather than glyphs. You can cover or black out areas and use every page-level operation. There is no OCR step.
Does this work on password-protected PDFs?: No. Encrypted files are detected when you open them and rejected with an explanation, rather than being partially loaded.
Is this the same as redaction?: It removes the text, which is the substance of redaction, but a compliance workflow usually also demands an audit trail and metadata scrubbing that this tool does not provide. For regulated redaction, verify the output against your own requirements.