Extract pages from a PDF into a new file — in any order you choose

How to use the PDF page extractor

1. Drop a PDF onto the box or press Choose a file. If you are not sure of the page numbers, look first in the PDF viewer.
2. In Pages to keep, type the pages in the order you want them, for example 2, 4-6. The default is 1, the first page only.
3. Press Extract pages.
4. Check the summary, which says how many pages were kept, then download the new file. It is named after your file with -extract added, so report.pdf becomes report-extract.pdf.

What it does and when to use it

Long PDFs often contain one or two pages that someone actually needs. Typical cases:

• Sending one part of a document. Your landlord needs page 3 of a 40-page bank statement, or HR needs only the signed last page of a contract.
• Uploading to a form with a size or page limit. A visa or job portal asks for “the first two pages of your passport scan” or a single-page certificate.
• Building a handout. Pull pages 12–15 of a textbook PDF into a short study pack, or keep only the slides you will present.
• Removing private pages. Share a report without the appendix that lists salaries or phone numbers.

If you know which pages to throw away rather than which to keep, the PDF page deleter lets you click the unwanted pages instead. If you want every page but in a new order, the reorder tool is easier.

How it works

Everything happens in your browser with pdf-lib (the @cantoo/pdf-lib package). No server is involved, and this tool does not download any extra data.

1. The file is read into memory and opened. If it only has an owner password (restrictions), it is opened without one.
2. Your page list is turned into page numbers. Ranges are expanded, odd, even, last and all are understood, and any number outside the file stops the job with a clear message.
3. Repeats are removed, keeping the first time each page appears.
4. The document’s page list is rewritten to contain only those pages, in your order. Page settings that a PDF can inherit from a parent (such as page size, rotation and fonts) are copied onto each page first, so nothing changes when the page tree is rebuilt.
5. Every object that no longer belongs to a kept page or to the document itself is deleted, and the file is saved with compressed object streams.

Because the original document is edited rather than copied page by page, document-level parts such as the title, author and other metadata stay with the extracted file. Bookmarks that pointed to pages you did not keep may no longer lead anywhere.

Page list syntax

Commas, semicolons or spaces can separate the parts, and 3 to 5 or 3–5 work like 3-5.

Worked examples

These summaries come from running the tool on a 12-page A4 report called report.pdf.

A short section. Pages to keep 3-5 → summary 3 of 12 pages kept (3-5). The new report-extract.pdf has 3 pages, starting with the original page 3.

Last pages first. 12-10, 1 → 4 of 12 pages kept (1, 10-12). The file holds pages 12, 11, 10 and then 1. This is useful for putting a signed final page in front of a cover page.

Every other page. odd → 6 of 12 pages kept (1, 3, 5, 7, 9, 11). Handy when a duplex scan has put blank backs on the even pages.

A typing slip. 14 → Page 14 is out of range — this PDF has 12 pages. Nothing is saved, so fix the number and press Extract pages again.

Limits and tips

• One output file per run. For several files, use the PDF splitter or run the extractor again with another list.
• No repeats. A page listed twice appears once.
• Big files need memory. The whole PDF is read into the browser tab. Files of hundreds of megabytes can make the tab run out of memory on phones.
• Page numbers are physical pages. If the document prints “page 1” on its fifth sheet, type 5, not 1.
• Shrink the result if needed. Extracted pages keep their full image quality, so a page from a scan can still be large. Run it through the PDF compressor or a fixed target such as compress PDF to 100 KB for upload limits.
• Keep your original. The tool never changes the file on your disk, but it is good practice to keep the full version.