Extract Pages from a PDF
Keep only the pages you want. Click thumbnails to select them, and the exported file contains those pages and nothing else.
Click the thumbnails you want to keep. Anything you do not select is dropped from the export.
How to extract specific pages
- Drop the PDF into the dashed box and wait for thumbnails. Long documents render as you scroll — you do not have to wait for the whole file.
- Click the thumbnails you want to keep. A blue "Keep" badge appears in the corner; click again to unselect.
- Press Export selected pages. The download contains only those pages, in their original order.
Why click instead of typing page numbers
Typing "1,3,5-8" looks efficient, but it hides an assumption: you already know which page is page 8. In practice people navigate a PDF by looking — "the form with the stamp on it". Thumbnails let you recognize the content and select it directly, which removes the scroll → note the number → scroll back → type it → discover it was wrong loop entirely.
With a few hundred pages, scrolling roughly to the right place and clicking beats guessing numbers in a text field every time.
Extract or delete — which one?
The output is equivalent; only the effort differs. The deciding factor is how many pages you are touching, relative to the whole document:
- 3 pages wanted out of 100 → extract, three clicks.
- 3 pages to remove out of 100 → delete, three clicks.
Choosing the wrong one costs you clicks, nothing else. There is no structural difference between the two outputs.
The page numbers will not match, and that is expected
People often assume something broke: they extracted pages 12 to 15, and the result opens showing "page 1, page 2…" while the footer of each page still says 12, 13, 14, 15.
Nothing broke. The number your reader displays is the page's position inside the file, and in the new file the first selected page is page 1. The number printed at the bottom is part of the page content — text drawn on the sheet, unrelated to document structure. To make the two agree, stamp fresh numbers with the page-numbering tool.
What it does technically
After parsing, the tool evaluates the keep-state of every page and removes the unselected nodes from the page tree — the same operation delete uses, with the condition inverted. Surviving page objects are untouched: no decoding, no redrawing.
Document-level information (title, author, outline) is therefore preserved, while page-level content (annotations, filled form values, in-page links) travels with the page itself. If you are sending the result outside your organisation, remember that comments can contain notes you would rather not share — the metadata cleaner cannot remove annotations, so delete them in a reader first.