Remove Blank Pages from PDF
Finds the empty backs and separator sheets in a scan, and lets you check before deleting.
What counts as blank
Your PDF is ready
Pages
No page is shown here. Switch to All pages, or raise the sensitivity.
About the Remove Blank Pages from PDF
A duplex scan of one-sided paper comes back with an empty page after every sheet, and batch scanners add separator pages. This tool draws every page of your PDF, measures how much ink is on it and shows you the ones that look blank — with a thumbnail, the share of the page that carries ink and why it was flagged. Nothing is removed until you press the button, and you can untick any page first.
Scanned “blank” pages are never white: they carry sensor noise, dust, a grey cast and often a black band at one edge. So the measurement ignores a band around the edges, counts only samples that are clearly darker than the paper — whatever its tone — and needs two dark samples in the same small tile before it counts anything — which drops single-pixel specks. A separate option treats a page whose only mark sits in the header or footer (a page number) as blank too.
How to use it
- Choose the scanned PDF (drag it onto the box, or press Choose a PDF). Every page is drawn and measured in your browser; a long scan takes a moment.
- Set How empty a page must be: “Scanned blank pages” suits most scans. Move it up to catch pages with a stray mark or a fold line, down to remove only completely empty pages. The list updates at once — the pages are measured only once.
- Tick Treat a page number as blank if your empty backs still carry a running header or footer.
- Look through the pages shown under Look blank: each one has its thumbnail and how much ink it carries. Untick anything you want to keep, or switch to All pages and tick a page the measurement missed.
- Press Remove the ticked pages, then Download PDF. The pages that stay keep their text, pictures, links and bookmarks; your original file is not changed.
Examples
48-page PDF from an automatic document feeder, every second page empty
24 pages flagged as “completely empty” (0 % ink) at the recommended sensitivity; the result has 24 pages.
36-page scan; the empty backs still print “Page 7 of 18” in the footer
With **Treat a page number as blank** on, those pages are flagged as “Only a page number” (about 0.3 % ink, all of it in the footer band).
Without that option they are kept, because ink is ink — the tool never guesses that a mark is “unimportant”.
A nearly empty page carrying a 3 cm received-stamp (about 1 % of the page)
Kept at the recommended sensitivity; flagged as “Nearly empty” only at the loosest setting, where the list shows its thumbnail so you can untick it.
Common uses
- Cleaning up a double-sided scan of one-sided paper before filing or sending it.
- Removing the separator sheets a batch scanner inserts between documents.
- Tidying a PDF printed from a program that adds an empty page after each section.
- Checking how many pages of a scan are actually empty before deciding to rescan.
How a page is measured
Each page is drawn with pdf.js (the PDF engine in Firefox) at 300 pixels wide — enough to see a line of text, small enough to measure hundreds of pages in a browser tab. Then:
- a band of 3 % around the edges is ignored, so scanner borders, punch holes and the shadow of the lid do not count;
- a sample counts as ink only when it is clearly darker than the paper — white paper, or the page’s own tone when it is darker (a grey cast, coloured or cream paper, a photographed page), which is taken from the page itself — so the paper never counts as ink;
- the page is divided into small tiles and a tile counts only when at least two of its samples are ink, which drops the single-pixel specks that scanners produce;
- the sensitivity then applies to the ink in the middle of the page (everything but the top and bottom 12 %), and ink in the header or footer band keeps the page — until you tick “Treat a page number as blank”, which allows a small amount of ink there.
What the list shows is exactly the number the decision uses, so you can always check it yourself.
Nothing is guessed about your document
The tool never decides on its own that a page is unimportant. It measures ink, you look at the thumbnails, and only the pages you leave ticked are removed. Pages that could not be drawn (a damaged page, or an image format your browser cannot decode) are listed as such and are never removed.
Removing pages rebuilds the file from the pages that stay, which keeps their content, links, form fields, layers and bookmarks; bookmarks that pointed at a removed page are dropped, because their target is gone. Some document-wide extras are not carried into the new file: page labels (i, ii, 1, 2 …), files attached to the document, the accessibility tags and the XMP metadata.
Text-only PDFs and OCR
The measurement works on what a reader shows, not on the text layer, so it works the same for scans, for PDFs exported from Word and for pages that hold only a picture. A page that carries invisible OCR text and nothing else counts as blank, which is usually what you want after scanning.
If you need the text of a scan as well, run OCR PDF first — the invisible text layer it adds does not change what this tool measures.
Limitations
- Pages are measured at 300 pixels wide. A single hairline or a very faint pencil mark can fall below that; check the thumbnails, and drop the sensitivity if you want only completely empty pages removed.
- At most 2,000 pages are measured in one go.
- A password-protected or encrypted PDF cannot be read here. Remove the protection with Unlock PDF first.
- Removing pages rewrites the file, so a digital signature in it will show as invalid, and bookmarks that pointed at a removed page are dropped.
- The new file does not keep the document’s page labels, attached files, accessibility tags or XMP metadata.
- Ink is ink: a page with a running header and nothing else is only flagged when you switch on “Treat a page number as blank”.
- Very large files need plenty of memory; on a phone close other tabs first.
Privacy
Everything happens in your browser. What you enter or open here is not uploaded or stored by MySmartCoPilot.
Frequently asked questions
Is my scan uploaded?
No. The pages are drawn with pdf.js and the new file is written with pdf-lib, both in your browser. Nothing is sent to MySmartCoPilot or anyone else, so you can clean up scans of confidential paperwork.
Why is a page with just a page number not flagged by default?
Because it is not empty, and in many documents those pages matter (a chapter divider, for example). Switch on Treat a page number as blank and such pages are flagged when all their ink sits in the header or footer band.
Will it delete a page that has a faint signature or a stamp?
Not at the recommended sensitivity: a stamp or a signature covers much more of the page than scanner noise. At the loosest setting such pages can be flagged, which is why every flagged page is shown with its thumbnail and its ink share before anything is removed.
Does it keep my bookmarks and links?
Yes, for every page that stays. The file is rebuilt from the pages you keep, with their text, pictures, links, form fields and bookmarks; only bookmarks whose target page was removed disappear. Page labels, files attached to the document, accessibility tags and XMP metadata are not carried over.
Can I use this on a PDF that is not a scan?
Yes. The measurement looks at what a reader shows, so an empty page from a word processor or a print-to-PDF job is found just as easily.
What if I change my mind?
Your original file is never changed — the tool writes a new copy. Untick pages before removing them, or press Start over and choose the file again.