Blank pages accumulate for structural reasons. A duplex scanner records the back of every sheet whether or not anything was printed on it, so a twenty-page single-sided document becomes forty pages, half of them empty. Chapter breaks in a printed book leave blank versos. Fax and scan software inserts separator sheets.
This tool finds and removes them automatically. Each page is rendered at low resolution and sampled pixel by pixel, counting anything darker than near-white as ink. A page is kept if more than 0.4% of the sampled pixels contain ink, and discarded if not. That threshold is deliberately low, so a page carrying only a page number or a thin footer rule is correctly recognised as not blank.
Detection and rebuilding both happen in your browser. Scanned personnel files or medical records are cleaned up without a copy reaching a server.
-edited suffix. Compare the page count against the original before discarding it.| Step | Detail |
|---|---|
| Render | The page is drawn at up to 320 pixels on its longest side |
| Sample | Every fourth pixel is examined, which is fast and statistically sufficient |
| Ink test | Any pixel with a channel below 235 of 255 counts as ink |
| Threshold | The page is kept if more than 0.4% of sampled pixels contain ink |
| Rebuild | Kept pages are copied losslessly — text and vectors intact |
Scans are the awkward case. A blank sheet run through a scanner is rarely pure white: it picks up paper texture, dust, a grey shadow along the binding edge, or faint show-through from the printing on the other side. If any of that pushes the page above the ink threshold, the page is kept and you will still see it in the output.
The opposite error is rarer but possible. A page containing only a very faint watermark, a light grey graphic, or a single line of pale text may fall below the threshold and be removed. This is why the tool reports how many pages it took out — if the number is higher than you expected, check the result against the original before discarding it.
The page is rendered small, every fourth pixel is sampled, and any pixel darker than near-white counts as ink. If less than 0.4% of the sampled pixels contain ink, the page is treated as blank.
No. The threshold is set low precisely so that a page number, a footer rule or a header line is enough to keep a page. Only genuinely empty pages are removed.
Scanned blanks are often not truly white — paper texture, dust, edge shadows and show-through from the reverse side all register as ink. Those pages exceed the threshold and are kept. Remove them by hand with Delete PDF Pages.
It is possible for a page with only very faint content to fall below the threshold. The tool reports how many pages it removed; check that number and compare against your original before deleting it.
No. Kept pages are copied as complete PDF objects, so text stays selectable and vector graphics stay sharp.
No. The rendering, analysis and rebuild all happen inside your browser.