Why was a page I consider blank kept?

It carries more ink than the threshold allows — a header, a page number, a faint stamp, or heavy scanner grain. The panel shows the measurement for every page it judged, so lower the threshold until that page appears in the list.

Why was a page with writing on it removed?

Almost always a very sparse page — a single line, or a signature on an otherwise empty sheet. Lower the threshold and it will be kept. Nothing is deleted in any case: the page is still in the rail, and one click restores it.

Does it work on a PDF that is not a scan?

Yes, for genuinely empty pages — those are found exactly rather than measured. Only pages that look like a scan get the pixel measurement, because that is the only kind of page that needs it.

Is my scan uploaded to measure it?

No. The pages are rendered and measured by your own browser, which is the only place the file ever exists. There is no server in this at all.

Remove blank pages from a PDF

How it works

01Open the scanned PDF.
02The panel lists every page it considers blank, with how much ink each one carries.
03Nudge the threshold if it has been too eager or too cautious.
04Export.

A scanned blank sheet is the hard case

Most tools test whether a page contains any objects. That finds an empty page in a document assembled on a computer, and never finds an empty page in a scan — because a scanned blank sheet is a photograph of white paper, and as far as the file is concerned that is a full-page image like any other. So scanned pages are measured from their pixels instead: the page is rendered small and the ink is counted. Both kinds of blank page are found.

It shows you the reading, not just a verdict

Any threshold has edge cases — a page with a single line on it, a chapter divider with one word. So the panel lists each candidate page with the percentage of ink it measured, before anything is removed. A page that should not be in the list is visible as a number that looks wrong, rather than as a page you notice missing a week later.

Scanner shadows do not count as ink

A flatbed leaves a dark band down the edge where the lid does not close flush, and a sheet fed askew leaves a black wedge in the corner. To a naive threshold both are ink, which is why some tools never find a blank page in a real scan at all. A small border is excluded from the measurement, so the shadow is ignored and a line of text near the top of the page still counts.

Nothing is deleted

Blank pages are hidden, not removed. They stay in the page rail, greyed out, and one click brings any of them back — so if the threshold took a page you wanted, you can see it and restore it rather than starting again. The export contains what the rail shows.

Splitting knows about this too

The same measurement drives splitting at blank pages, which is the other half of the scanning workflow: put a blank sheet between documents, scan the whole stack in one pass, and get one file per document back. Both tools agree about what counts as empty.

Questions

Why was a page I consider blank kept?: It carries more ink than the threshold allows — a header, a page number, a faint stamp, or heavy scanner grain. The panel shows the measurement for every page it judged, so lower the threshold until that page appears in the list.
Why was a page with writing on it removed?: Almost always a very sparse page — a single line, or a signature on an otherwise empty sheet. Lower the threshold and it will be kept. Nothing is deleted in any case: the page is still in the rail, and one click restores it.
Does it work on a PDF that is not a scan?: Yes, for genuinely empty pages — those are found exactly rather than measured. Only pages that look like a scan get the pixel measurement, because that is the only kind of page that needs it.
Is my scan uploaded to measure it?: No. The pages are rendered and measured by your own browser, which is the only place the file ever exists. There is no server in this at all.