Extract pages from a PDF
Need only page 4 of a statement, or the signed page of a contract? Choose those pages and download them as a separate PDF — in the order you want, with the original left untouched. Your file is never uploaded.
Drop a PDF here or click to choose
Pages pulled out on your device — never uploaded
How to extract pages from a PDF
- Drop the PDF above or click the box to choose it.
- Enter a From/To range under Page range, or switch to Pick pages and click the previews you need. To change the order, type the list under Advanced: type a page list.
- Press Extract pages. The result shows how many of the original pages went into the new file and how big it is.
- Download it. Nothing is changed in the file you started with.
Extract or delete? Count the shorter list
Both tools end with the same kind of file — the original minus some pages — so pick the one that needs fewer numbers typed or clicked.
| What you want to end up with | Use |
|---|---|
| 3 pages out of a 200-page report | Extract (type 41-43) |
| Everything except a cover and a blank last page | Delete pages (type 1, 200) |
| The same pages, but in a new order | Extract, with a typed list such as 5, 1-4 |
| Several separate files from one long PDF | Split a PDF |
| Reorder, rotate and remove by dragging previews | Organize PDF |
How big is the extracted PDF?
Smaller than the original, but not always in proportion to the page count. We ran the extraction this page performs — the same qpdf build and the same kind of page list — on three generated test documents, and read every result back with pdf.js to confirm which pages came out.
| Test document | Pages kept | Original | Extract |
|---|---|---|---|
| Plain text, 296 pages | 12–14 (3 pages) | 60.3 KB | 1.3 KB |
| Scanned sheets, 10 pages | 1-3 (3 pages) | 2,958.0 KB | 888.3 KB (−70%) |
| Report with a letterhead on every page, 20 pages | 5 (1 page) | 53.8 KB | 28.2 KB (−48%) |
The scans behave the way you would expect: three of ten pages, about three tenths of the size, and the kept scans’ JPEG data came out byte-for-byte identical to the original. The report does not: a single page still weighs more than half the file, because the letterhead image is stored once and travels with any page that shows it. If the extract has to fit an email limit, run Compress PDF on it afterwards.
Page lists, order and repeats
| You type | New PDF, in this order |
|---|---|
4 | 4 |
1, 5-7 | 1, 5, 6, 7 |
5, 1-2 | 5, 1, 2 |
1-3, 1 | 1, 2, 3, then page 1 again |
Numbers always refer to the original document. A number beyond its
last page, or a backwards range such as 7-5, is refused
with a message rather than quietly ignored.
Why extract in the browser?
Pulling out one page is usually about holding the rest back. The landlord asks for proof of income, not the whole bank statement; the visa form wants one certificate from a scanned bundle; the client gets the signature page, not the internal annex. Uploading the full document to a website to isolate that page gives a stranger exactly the pages you meant to keep to yourself.
Here the copy happens in your browser’s memory and the download is created on your device. The PDF is not sent anywhere for processing, there is no server-side size quota, and very large files are limited only by the memory your device can spare.
Frequently asked questions
Is the PDF uploaded to pull the pages out?
No. qpdf, an open-source PDF engine, runs in this tab and builds the new file from the copy your browser already read from disk. Loading the page, fetching the engine and showing ads are ordinary network requests; the document is never sent along with any of them.
Can someone recover the pages I did not extract?
No, because they are not in the file. The output is assembled from your list alone. In our check, a scanned page that was left out had none of its image data in the new PDF, and the same held for left-out text pages. A viewer that only hides pages, or a crop, is a different thing: there the content is still inside.
How do I put the pages in a different order?
Open “Advanced: type a page list” and write them in the order you want. “5, 1-2” gives a three-page PDF that starts with the old page 5; we confirmed the order by reading the text back from the output. A page typed twice appears twice. Clicking previews or using From/To always gives the pages in their original order.
Does “12–14” with a long dash work?
Yes. A page list pasted from Word, a PDF or a chat app often has an en dash (–), an em dash (—) or a minus sign (−) where a hyphen should be. All three count as a range, so “12–14” extracts pages 12, 13 and 14.
Why is a one-page extract still half the size of the whole file?
Because some things are stored once for the whole document — a letterhead image, a logo, embedded fonts — and every page that shows them needs them. Our 20-page test report weighed 53.8 KB; its page 5 alone was 28.2 KB, almost all of it the letterhead. Documents where each page carries its own content, like scans, shrink in proportion instead.
The PDF says it cannot be edited. Can I still extract pages?
If it opens without a password, yes. Such a file has an owner lock only; it is removed on your device so the pages can be copied, the tool says so, and the new PDF does not carry it. If the PDF needs a password to open at all, this site cannot help: there is no password field here, by design.
What will the new file be called?
A single range names itself after the pages: report.pdf with “1-3” becomes report-pages-1-3.pdf, and “4” gives report-page-4.pdf. Anything more mixed, such as “1, 5-7”, is saved as report-extracted.pdf. The original file is never changed.