Extract or Delete PDF Pages
Keep a subset of pages, or remove unwanted pages, and save the result as a new PDF. Everything happens on your device.
This is the inverse of splitting. Instead of producing many files, it produces one clean document: either only the pages you listed, or everything except the pages you listed. It is the fastest way to strip blank scans, remove an appendix, or drop a redacted section before sending a file on.
The pages you keep are copied exactly as they are, including annotations, form fields and embedded fonts. Pages you drop simply do not appear in the output document.
Deleting a page is not the same as redacting it
These two jobs are constantly confused and the difference matters. Deleting a page removes it from the output document: the text on it is gone, because the page object itself is gone. Drawing a black box over a paragraph does the opposite. The rectangle is painted on top, and the text underneath is still in the file, still selectable, and still recoverable by anyone who selects it, copies it with a text tool, or opens the file in an editor. If the goal is to stop someone reading a name, an account number or a clause, delete the page, or cover the text properly at the content level. A black rectangle over live text is a curtain, not a wall.
Where PDF page numbers disagree with printed ones
This is the single biggest source of mistakes when trimming a report or a contract. The tool counts pages physically, starting from the first page of the file, exactly as a reader scrolls through it. Printed documents very often start their own numbering later than that: a cover, a title page and a contents list are usually unnumbered, and the body then begins at printed page 1. In a file with four unnumbered front pages, "delete page 12" therefore means printed page 8, not printed page 12.
Three ways to get this right, in rough order of how well they hold up:
- Count from the first page of the file using the thumbnails in the file list, ignoring the numbers printed on the pages.
- If the document has front matter with its own numbering, work out the offset once and apply it for the whole document.
- If the document carries Bates numbers or a continuous stamp, those are normally applied from the first physical page and make a reliable index.
Trimming blank scans out of a batch
A duplex feeder produces a blank page every time it meets the back of a single-sided sheet, and a 200-page scan can easily carry thirty of them. Use delete mode and list the blanks. It is faster and safer than listing the pages you want to keep: a mistake in keep mode silently drops real content, while a mistake in delete mode usually leaves an obvious blank page behind for you to notice.
What goes with a deleted page, and what stays
- Annotations on a deleted page go with it. Annotations on kept pages are preserved.
- Outlines and bookmarks are not rewritten. An entry that pointed at a deleted page can end up pointing at the wrong page, so check the bookmarks panel before sending a trimmed file to anyone who navigates by it.
- The file usually gets smaller, but not proportionally. Objects shared between pages — an embedded font, a logo placed on every page — are retained as long as any remaining page still refers to them.
- Nothing is written back to your original. Only a new document is produced, so a page list can be tried as many times as it takes.
How to extract or delete PDF pages
- Add your PDFDrop the document you want to trim.
- Choose keep or deleteKeep mode outputs only the listed pages. Delete mode outputs everything except them.
- List the pagesEnter ranges such as 1-2, 4, 9-11 using 1-based page numbers.
- Save the new PDFThe result is generated locally and downloaded immediately.
Frequently asked questions
What is the difference between extract and split?
Can I delete several scattered pages?
Are page numbers 1-based?
Does this preserve form fields on kept pages?
Can I undo the result?
Related tools
Related guides
Every tool on this site runs locally. Read the no upload policy to see how to verify it yourself.