Extraction is a copy operation, not a rendering one. The page objects are lifted out of the source file and written into a new document along with the fonts and images they reference, so the result stays selectable, searchable and printable at full resolution. Nothing is flattened into a picture, which is what separates a real extraction from screenshotting pages one at a time.
The work happens on your own machine. The file is read with the browser File API, rebuilt in memory and handed back as a download, so a signed contract never travels across a network. There is no queue and no account.
When you should extract PDF pages rather than split the file
Splitting and extracting sound similar and solve different problems. Splitting cuts a document into every one of its parts, which is what you want when a scanner has stapled twenty invoices into one file. Extracting pulls out a specific subset and throws the rest away, which is what you want when someone asked for the appendix and nothing else.
The practical test is whether you care about the pages you are not keeping. If they belong in their own files, split. Otherwise, extract.
- Send one chapter of a long report without attaching the whole report.
- Pull the signature pages out of a contract for countersigning.
- Keep only the pages of a bank statement that cover the period being queried.
- Lift a single form out of a scanned bundle before filling it in.
- Cut a 400 page manual down to the twelve pages a colleague actually needs.
How to write a pdf page range
The range box accepts single page numbers, spans and any mixture of the two separated by commas. Pages are numbered from 1, matching what your PDF reader shows in its page field rather than a zero based index.
Order matters. The pages come out in the order you list them, so 5, 1, 3 produces a three page document that starts with page five. That makes the range box a quick reordering tool as well as a filter. Repeated numbers are dropped by default, but you can turn that off when you deliberately want a page to appear more than once.
What survives when you extract pages from pdf files
Everything that belongs to the page comes with it. Text stays as text with its embedded font subsets, images keep their original compression and resolution, and vector drawings stay as vectors. Link annotations that point inside the extracted set continue to work, and page dimensions and rotation are carried across unchanged.
Some document level features do not survive, because they describe the whole file rather than any single page. Bookmarks, the outline and links that pointed to pages you dropped cannot be preserved in a subset that no longer contains their targets. Metadata such as the title and author is not copied either, since the new file is a new document.
If those features matter, delete the unwanted pages from a copy of the original instead. That keeps the document intact and only removes what you named.
How this extract pdf pages tool works in your browser
Two libraries do the work and neither of them needs a server. A PDF parser reads the file structure and renders the thumbnail strip so you can confirm the range visually, and a PDF writer copies the chosen page objects, resolves everything they reference and serialises a new document.
Because the compute is yours, the limits are yours too. A 300 megabyte scanned archive is bounded only by the memory your browser tab can claim, not by an upload cap. You can verify the privacy claim directly: open the network panel in your developer tools and run an extraction while watching it, and no request carries your file.
The one thing a browser cannot do is guess a password. An encrypted PDF is reported with a clear message rather than failing silently, so you know to remove the protection in the application that created it first.
Using the pdf page extractor on long documents
Thumbnails are rendered for the first two dozen pages only. Rendering every page of a long document is the slow part of opening it, and for a 500 page file that would cost several seconds before you could type anything. Any page number up to the real page count still works in the range box, whether or not a thumbnail exists for it.
Extraction itself is close to instant regardless of length, because copying page objects is structural rather than visual work. The output is usually far smaller than the source, though not proportionally: a page drags in the parts of every font it uses, so a one page extract carries a small fixed overhead.
| You type | You get | Notes |
|---|---|---|
| 1-3 | Pages 1, 2 and 3 | A span, inclusive at both ends |
| 4 | Page 4 only | A single page |
| 1-3, 8, 12-14 | Seven pages | Spans and singles can be mixed |
| 9, 2, 5 | Three pages in that order | Order follows what you typed |
| 2, 2, 2 | One page, or three | Depends on the repeated pages toggle |
How to extract PDF pages
- 1
Open your PDF
Drag a PDF onto the drop area or click to browse. The page count appears immediately and a thumbnail strip renders below it.
- 2
Type the pages you want
Enter single pages, ranges or both, such as 1-3, 8, 12-14. Clicking a thumbnail appends that page number to the box.
- 3
Decide about repeats
Leave the duplicate option on for a normal extraction. Switch it off when you want a page to appear more than once in the output.
- 4
Extract and download
Press extract and download the new PDF. The original file on your disk is never modified by the tool.
Related tools worth bookmarking
Sources and further reading
- ISO 32000-2: Portable Document FormatThe published PDF specification, which defines the page tree structure that makes page level extraction possible.iso.org
- MDN: Using files from web applicationsThe browser file APIs this tool relies on to read a document locally without any upload.developer.mozilla.org
Frequently asked questions
Common questions about the extract pdf pages.
No. The PDF is read, parsed and rewritten by JavaScript running inside your own browser tab, and the result is produced as a local download. No copy of the document is transmitted, stored or seen by anyone else at any point in the process.
No. Pages are copied as objects rather than re rendered, so text remains selectable text, images keep their original pixels and compression, and vector graphics stay sharp at any zoom level. The output is visually identical to the pages in the source file.
Yes. Pages come out in the order you type them, so entering 9, 2, 5 produces a document in that sequence. This makes the range box a fast way to reorder a small selection while you are already extracting it.
Links between pages that are both inside your selection keep working. Bookmarks, the document outline and links pointing at pages you did not extract are not carried over, because their targets no longer exist in the new document.
Encrypted PDFs cannot be decoded without the password, and the tool does not ask for one. Open the file in a reader that has the password, save an unprotected copy, and extract from that copy instead. The error message tells you when this is the cause.
The tool imposes none. The real limit is how much memory your browser tab can allocate, which on a current laptop comfortably covers documents of several hundred megabytes. Very large scans are slower to preview than to extract.