Extract pages from a PDF — privately, in your browser
You rarely need the whole document. A lender wants two pages of a tax return, not all forty; a colleague needs the appendix, not the report; a court filing requires one exhibit out of a long scan. This tool builds a new PDF containing only the pages you name, using 1-based page numbers with comma-separated values and ranges — typing 1,3-5 keeps pages 1, 3, 4, and 5. Everything runs in your browser through pdf-lib, pure JavaScript with no server involved. That is deliberate: the documents people extract pages from are disproportionately sensitive — tax filings, medical charts, contracts — and sending them to a stranger’s server to save thirty seconds is a bad trade. Your original file is never modified; you get a fresh PDF with just the pages you asked for.
How it works
- Drop your PDF below or click to select it.
- Type the pages to keep in the Pages box, like 2,5-8,11 (page numbers start at 1).
- Download the new PDF containing only those pages.
Frequently asked questions
What page range formats are accepted?
Comma-separated page numbers and ranges, counted from 1. Examples: 3 (one page), 1,4,7 (three individual pages), 2-6 (a run of pages), or any mix like 1,3-5,9.
Does my document leave my computer?
No. Extraction is performed by pdf-lib running as JavaScript inside the page you are looking at — there is no upload step at all. Given how often the file in question is a tax return, a medical record, or a signed agreement, keeping it local is not a nice-to-have; it is the reason this tool exists.
Does it work on password-protected PDFs?
Sometimes. The tool attempts to open encrypted files by ignoring the encryption, which works for PDFs that merely restrict printing or editing. Heavily protected files — ones that demand a password just to open — may fail. If that happens, unlock the file with its password in a PDF viewer first, then extract the pages here.
Is the original PDF changed?
Never. The tool reads your file and writes a brand-new PDF containing the requested pages; the source document on your disk is untouched. If you asked for the wrong range, just run it again with a corrected one.