How to remove watermarks and page numbers when converting a PDF to an ebook (remove watermark pdf to epub)

Last updated: 7 November 2026

Page numbers and running headers are usually easy to drop when you convert a PDF to EPUB, because they repeat in the same spot on every page and a converter can recognise that pattern. Watermarks are harder. A text watermark can often be filtered out, but one baked into a page image can't be separated cleanly, so the best fix is a cleaner source file.

This guide covers how to tell which kind of clutter you have, what to do about each kind, and how to check the result on your reader.

Why do page numbers, headers and watermarks end up in my EPUB?

They end up there because a PDF describes a printed page, not a flow of text. Page numbers, running titles and watermarks are all just marks at fixed positions, so a converter that doesn't recognise them treats them as part of the story.

In a PDF, "page 47" is text placed at the bottom of a 47th page. When an EPUB is made, that page boundary disappears, because the text reflows to fit your screen. If the converter doesn't strip the number, you get a stray "47" in the middle of a sentence. You see the same thing with a chapter title repeated every page, or a line like "Downloaded from…" interrupting paragraphs.

Typical clutter falls into four groups:

• Page numbers: a lone digit or roman numeral, often with dashes ("– 47 –").
• Running headers and footers: the book title, author name or current chapter, repeated on each page.
• Watermarks: diagonal or centred text such as "DRAFT" or "Confidential", or a site logo behind the text.
• Ads and promo lines: download-site banners, URLs or "scan by…" credits.

Which kind of clutter do I have? A 2-minute check

Open the PDF and try to select the clutter with your cursor. What happens tells you how fixable it is.

• You can select and copy it as text. It's live text. A converter that detects repeating elements can usually remove it.
• You can't select it, but you can select the body text. It's probably an image or vector overlay, which is common with watermarks. It may be ignored when only the text is extracted, but any image-based conversion will keep it.
• You can't select anything. The whole PDF is a scan. See our guide on how to tell whether a PDF is scanned or text, because the clutter now sits inside page images and needs OCR first.

How do I remove page numbers and headers when converting a PDF to EPUB?

Use a converter that detects text repeating in the same position on every page and drops it, then check the chapter joins. If it misses some, delete the leftovers afterwards with a find-and-replace pass.

Let the converter do the first pass

Our own demo page uses an illustrated public-domain edition of Alice's Adventures in Wonderland to show the idea. Its before-and-after list says that in the EPUB, printed page numbers leave the story text and watermarks and ads are stripped away automatically. That's the behaviour you want, and it's what the site describes for Convert2EPUB. Results still depend on the file. A clean novel is the easy case, and a messy layout is harder.

On our converter, the practical steps are:

• Upload the PDF (up to 100 MB).
• Choose Standard for a text PDF with a simple layout like a novel. It's free and uses a rule-based engine on the PDF's own text.
• Leave the mode on Auto or choose Reflowable, since fixed layout keeps the page as it was, clutter included.
• Pick your target device, then convert and open the file on your reader.

Check the likely trouble spots

After converting, don't read the whole book to check it. Look at these places:

• Page boundaries mid-sentence. Search for a lone number between two lowercase words.
• The first line of each chapter. Running headers often attach themselves there.
• The front matter. Roman numeral page numbers (i, ii, iii) are the ones most often left behind.
• Footnotes. A real footnote number looks like a page number. A converter has to tell them apart, and mistakes happen both ways.

If a few strays remain, an EPUB editor such as Sigil, or a text editor with regular-expression search, can delete them. A pattern that matches a line containing only digits finds most leftover page numbers, but it will also match a line that's genuinely a number, such as a numbered list, so review each match before replacing.

When the table of contents is affected

Running headers can confuse chapter detection, because a repeated title looks like a chapter heading on every page. If your contents list shows the same entry over and over, see our guide to how the PDF to EPUB table of contents is built and how to fix a missing one.

Can I remove a watermark from a PDF before converting it?

Sometimes, but whether you should depends on who added it. If the watermark is your own (a "DRAFT" stamp on your manuscript, say), remove it at the source. If someone else applied it, check you have the right to alter the file first.

Cases where removal is straightforward

• You have the original document. Re-export from Word, Google Docs or whatever produced the PDF, with the watermark switched off. This gives the cleanest result by far, and Word, RTF, TXT and ODT files can be converted straight to EPUB without going through PDF at all.
• The watermark is a separate text layer. Some PDF editors can delete it as an object.
• The "watermark" is a promo line from a download site. This is live text that a converter or a find-and-replace pass can remove.

Cases where you should stop and think

Watermarks aren't always clutter. Authors and retailers use them deliberately. BookFunnel's help documentation describes a hidden watermark and an optional visible watermark inserted into EPUB files, and says the author can switch the visible one off. Watermarks like these identify a purchased copy. Stripping one from a book you don't own, or in order to share it, may breach the licence or copyright law. We can't tell you what's lawful in your country. We'd only suggest converting files you own or have permission to change.

Also note that not every watermark is visible. Hidden watermarks are designed to survive format changes, so converting a file and having the mark vanish from the page isn't evidence that it's gone.

What if the PDF is scanned?

Scanned PDFs need OCR before anything else, and the clutter then has to be filtered out of the recognised text. Expect more cleanup than with a text PDF, because stray marks get read as characters.

A scan is a photo of each page, so a page number is a pixel pattern until OCR turns it into the characters "47". Convert2EPUB has a "Read scanned pages (OCR)" option for pages with no selectable text. It says the pages become real, resizable text and the illustrations stay. You choose the OCR language from a list that includes English, Turkish, German, French, Spanish, Italian, Portuguese, Dutch and Greek.

For tips specific to scans, see how to convert a scanned PDF to EPUB with OCR. Three habits help most:

• Use the highest-resolution scan you have.
• Choose the right OCR language, because a wrong one produces junk characters.
• Check the margins. Shadows from a book's spine and scanner edges often become stray punctuation.

Which conversion model for cluttered pages?

The site says Pro normally reads only the pages that need it and leaves clean text pages free, and that you can tick a box to read and pay for every page. New accounts get 30 free credits, according to the converter page. Try Standard first on a short sample, then move to Pro only if the clutter persists.

How do I check the result on my own e-reader?

Open the EPUB on the device you'll read it on, then change the font size twice. Clutter that was hidden at one size often shows up at another, because reflow moves page-break leftovers into the middle of paragraphs.

A short checklist:

• Open the first chapter, a middle chapter and the last page.
• Increase and decrease the font size and look for stray numbers.
• Check the table of contents links.
• Search for the book's title or the author's name. Hits inside body text suggest a leftover running header.
• Check that images such as illustrations are still in place.

If you read on a Kindle, the format question is separate. Our guide on sending an EPUB to your Kindle covers that. For why a raw PDF looks so cluttered on a small screen, see why a PDF is hard to read on a Kindle.

If you'd like to see how a file comes out, you can upload a PDF to Convert2EPUB and look at the result on your device before you rely on it.

Frequently asked questions

Does converting a PDF to EPUB automatically remove page numbers?

Not always. It depends on the converter. Tools that detect repeating elements at the same position on each page usually remove them, and simple layouts such as novels work best. Check the page joins and the front matter afterwards, because roman numerals are often missed.

Can I remove a watermark from a scanned PDF?

It's difficult, because the watermark is part of the page image. OCR may read it as text, or ignore it if it's faint. A cleaner scan or the original file is the more reliable route.

Is it legal to remove a watermark from an ebook?

That depends on whose watermark it is and the law where you live, and we can't give legal advice. Removing your own marks, such as a draft stamp on your manuscript, is fine. Watermarks added by a seller or publisher often identify your licensed copy, so check the terms before changing the file.

Why does my EPUB still show a chapter title above every page break?

It's a running header that the converter treated as body text. Delete the repeated lines with find-and-replace in an EPUB editor, or re-convert with a tool that detects repeated headers. Check the table of contents as well, since repeated titles can create duplicate entries.

Is it better to convert from Word instead of PDF?

Yes, if you have the original. A DOCX has real headings and no printed page furniture, so there's nothing to strip out. A PDF only records how the pages looked.