Remove Blank Pages from a PDF

How a page is judged. pdf.js draws each page to a greyscale picture (about 640 pixels on the long side). A pixel counts as ink if it is darker than a "white level". The share of ink pixels is the page\'s ink percentage. A page with no ink is blank, a page with an ink share below a small limit is nearly blank (dust, a scanner speck, a hole punch shadow), and everything else has content. Because the page is drawn, text, images and vector drawings count alike, and white text on a white page is correctly seen as empty.

Sensitivity. Three levels. Strict only treats a completely empty page as blank. Normal ignores dust and tiny specks (up to about 0.04% of the page). Lenient also accepts pages with a few marks (up to about 0.4%). The white level changes with them. Changing the level re-checks all pages instantly and re-ticks the suggestions.

Scanner borders. Dark edges are common on scans. By default 2% of the page edge is ignored; set it to 0 if page numbers or headers sit close to the edge.

False positives. A page with only very light content (pencil, a pale watermark, a faint stamp, thin grey lines) can look blank. Pages that are called blank or nearly blank but contain very pale marks carry a warning. Look at those thumbnails before you remove them.

Removing. The pages you tick are removed with pdf-lib from the current version of the file, and the rest stay in their original order with their content, links and bookmarks. At least one page must remain.

PDF files up to 80 MB and 500 pages, including scans. Password-protected files must be unlocked first. The result is a normal PDF with the other pages unchanged; the original file is not modified.

1. Choose a PDF. Each page is drawn once; large files take a moment.
2. Pick a sensitivity. Pages that look blank are ticked automatically.
3. Check the thumbnails, especially pages with the "very light marks" note, and tick or untick pages.
4. Press the remove button and download the new PDF.

• Cleaning up a scan of double-sided documents where the backs are empty
• Removing separator sheets between scanned batches
• Dropping empty trailing pages from an exported report
• Preparing a shorter file for email or printing
• Checking how many pages of a PDF are really empty

Detection looks at what is drawn, not at what the page means. Faint scans, pencil notes, light watermarks and white-on-white text can be called blank, and dust, shadows, show-through from the other side and scanner edges can make a blank page look like content. Always check the thumbnails; nothing is removed until you press the button.

Pages are examined at low resolution, so one very small mark can be missed on a large page. Only whole pages are removed; the tool does not clean marks off a page. Documents with very many pages take longer and use more memory.