Select the pages you need and export them into a new PDF from supported document files, then download your extracted PDF in seconds.
1-3, 5, 7-9 into the Page ranges field. Each comma-separated token becomes its own output PDF — 1-3 yields a 3-page file, 5 yields a 1-page file, and 7-9 yields another 3-page file. Whitespace around commas is ignored; numbers must fall within the source document's page count.A 400-page board deck, a multi-chapter contract, or a scanned bank statement bundle often contains exactly two or three pages you actually need to share. Re-saving the whole file wastes storage and email allowance, exposes confidential pages, and forces recipients to scroll past noise. Extracting only the pages that matter is the cleanest fix:
Need related tooling? Try Split PDF when you want to break a file into equal chunks, Merge PDF to combine the extracted pieces back together, or Compress PDF if the extracted pages are still heavy from embedded scans.
| Operation | What it does | Result | When to choose |
|---|---|---|---|
| Extract | Pulls specific pages or ranges into new PDFs | One file per range token (e.g., 1-3,5,7-9 -> three PDFs) |
You know exactly which pages you want and don't need the rest |
| Split (page by page) | Cuts every page into its own file | N single-page PDFs from an N-page source | Bursting a scanned bundle so each page can be re-named or routed individually |
| Split (by range size) | Divides into equal-size chunks | Multiple multi-page PDFs of the chosen size | Breaking a 500-page archive into 50-page chapters for upload caps |
| Reorder | Rearranges existing pages within one PDF | Same page count, new sequence | The pages are all relevant but currently in the wrong order |
| Delete pages | Removes specific pages, keeps the rest | One PDF minus the listed pages | The unwanted set is smaller than the wanted set |
Extract and "delete pages" are inverses of each other: extracting 1-3, 7-9 from a 10-page PDF produces the same content as deleting 4-6, 10 — pick whichever expression is shorter.
| You type | What you get | Notes |
|---|---|---|
5 |
One PDF with just page 5 | Single page, no hyphen needed |
1-3 |
One PDF with pages 1, 2, 3 | Hyphen denotes an inclusive range |
1-3, 5, 7-9 |
Three separate PDFs (3 pages + 1 page + 3 pages) | Commas separate independent ranges |
10-15, 30, 40-60 |
Three PDFs (6, 1, and 21 pages) | Ranges can be far apart in the source |
1, 1, 1 |
Three PDFs each containing page 1 | Duplicates are allowed |
1-1 |
Same as 1 |
Single-page range is legal |
5-3 |
Error / ignored | End must be greater than or equal to start |
The largest page number must not exceed the source document's last page; out-of-range tokens are skipped with a warning. Whitespace around commas and hyphens is optional. There is no "open-ended" syntax like 7-end — type the final page number explicitly.
Yes — that's exactly what Pages by multi-range is for. Enter 1, 5-8, 12 and you'll get three output PDFs: a 1-page PDF (cover), an 8-page PDF (chapter), and another 1-page PDF (back-matter). Each comma-separated token becomes its own file in the resulting ZIP.
It depends on where the bookmark or link points. PDF outline entries (the bookmark tree defined in ISO 32000) reference pages by indirect object — when an extracted page is carried over, bookmarks pointing into that page are preserved. Bookmarks pointing to pages that are not in the output (e.g., a "Chapter 5" bookmark when you only extracted Chapter 3) are dropped, because the destination no longer exists. Internal cross-references (TOC links, "see page 47" hyperlinks) behave the same way. External URL links inside extracted pages are always kept.
If the form widget physically sits on a page you extracted, the visible field comes with it. The underlying AcroForm dictionary that ties widgets to a field hierarchy is handled tool-by-tool — well-behaved extractors (pypdf's append, qpdf, Acrobat) rebuild it so the field stays fillable. Page-only copy operations (older pypdf add_page, some quick-and-dirty splitters) may leave the widget visible but unbound, which means the field renders but doesn't post a value when submitted. Test the extracted PDF in Acrobat or your target viewer before relying on it for live form workflows.
The PDF's physical page index always restarts at 1 — page 7 of the source becomes page 1 of the new file as far as Acrobat's page counter and the URL #page= anchor are concerned. Any page numbers printed inside the page content (headers, footers like "Page 47 of 200") are decorative text and do not change — they still say 47. If you need clean numbering, run the output through a header/footer editor afterwards.
Extract pulls specific pages or ranges you name. Split divides the entire source either page-by-page (every page becomes a file) or by range (equal chunks). Use Extract when you want a small subset and the rest is irrelevant; use Split when you want to burst the whole document. The xconvert tool exposes both via the Page by page, Pages by range, and Pages by multi-range modes on a single page.
Usually yes, but not always proportionally. Page content streams are independent, so removing pages does drop bytes — but shared resources like embedded fonts, images, and color profiles are written once at the PDF root and may be copied over even when only one page references them. A 100 MB PDF with one 50 MB embedded image used on every page extracted to a single page typically lands around 50-55 MB, not 1 MB. Run the result through Compress PDF if you need it smaller.
files are processed on our servers and deleted automatically after a few hours and are not retained on permanent storage. No account is required, no email is asked for, and there is no watermark on the output. If you're working with privileged or regulated material (legal, medical, financial), this is a meaningful difference versus tools that route the file through their servers indefinitely.
The browser-side flow is constrained by the memory available to your tab rather than a fixed server cap, so the practical ceiling depends on your device — modern desktops handle 500-700 MB PDFs reliably; phones and low-RAM laptops tap out earlier. If you hit a tab freeze on a very large scanned PDF, run Compress PDF first to bring it under ~200 MB, then extract.
You need to know the password. Owner-restricted PDFs (where editing is locked but the file opens normally) can sometimes be processed without it, but user-password PDFs that won't open at all in a viewer cannot be extracted — the page content streams are encrypted and unreadable until decrypted with the correct password. Use a separate unlock step first if you have legal authority to do so.