To extract pages from a PDF means typing page numbers like 5, 1-3 into a selection field and letting the tool copy those pages — in that exact order — into a brand-new PDF, while the source file is read locally in your browser without ever being uploaded. For a beginner, the workflow comes down to three actions: load the source PDF from your device, list the pages you want, and download the resulting file. The tool that solves this task is Extract PDF Pages, and it works entirely in your current tab so the source bytes never leave your computer. The new file contains only the pages you selected, in the order you typed them, and the rest of the original document is left behind. This guide walks through every step, the syntax for selecting pages, the limits to respect, and the trade-offs to know before you share or archive the result.
What Extracting Pages Does to a PDF
Extracting pages from a PDF is different from deleting or splitting them. Deleting removes pages from the same file, leaving you with a shorter version of the original. Splitting breaks a PDF into several smaller files, often one per page or one per range, so you end up with many files instead of one. Extracting pages builds a fresh, separate PDF that contains only the pages you choose, in the order you choose them, and leaves the source PDF exactly as it was on disk.
This distinction matters for beginners because the typical reasons people want to "pull out" pages — sharing a single chapter, sending a few signed pages, archiving a section of a long report, or submitting only the relevant forms — all involve handing the recipient a new file rather than editing the original in place. Extract PDF Pages is built for that task: you pick the source, list the pages, and download a second file that only contains those pages.
Limits and Requirements Before You Start
Before you open the tool, check that your PDF fits the boundaries it accepts. The source must be an unencrypted PDF no larger than 50 MiB (52,428,800 bytes) and it must contain 500 pages or fewer. A file at either limit is accepted; a file that goes over either limit is rejected rather than shortened. Password-protected files are not opened by bypassing encryption, so if your PDF requires a password to view, you need to remove the protection first using a separate tool.
Browser memory also plays a role. Because the new file is built in your current tab, very large or image-heavy source PDFs may behave differently on different devices. Closing other tabs can free memory if extraction feels slow. The output file can still be larger than you expect because selected page resources such as fonts and images are copied along with the page, so extracting one page from a resource-heavy document can retain a significant amount of data.
| Limit | Accepted value |
|---|---|
| Source file size | Up to 50 MiB (52,428,800 bytes) |
| Source page count | 1 to 500 pages |
| Selection field length | Up to 4,000 characters |
| Encryption | Unencrypted PDFs only |
| Form type | Dynamic XFA forms are not a supported preservation target |
How to Extract Pages from a PDF Step by Step
Once your PDF meets the limits above, the actual extraction is short. Open the Extract PDF Pages tool in your browser and follow these steps.
- Choose an unencrypted PDF no larger than 50 MiB and with no more than 500 pages from your device. The browser validates the file and reads its bytes locally.
- Enter one-based page numbers or inclusive ascending ranges in the output order you want, for example 5, 1-3. Token order is significant: 5, 2-3 produces page 5 followed by pages 2 and 3.
- Run the extraction. The parser keeps the first occurrence of each unique page in your list and copies that exact set into a fresh PDF document.
- Review the duplicate-page notice if one appears, then download the new PDF. The source file, the selection, and the output are not uploaded.
Changing the page expression immediately releases the previous result, and running extraction again replaces the previous output URL. Choosing a different source file clears the old bytes, the old selection, the source link, the previous result, and any error message. The tool uses guards that prevent an older asynchronous load or save from overwriting newer state, so you can re-run a job without worrying about a stale result sneaking in.
Page Selection Syntax and Order Rules
Page numbers are one-based, which means page 1 is the first page shown by any PDF viewer. Tokens are separated by commas or whitespace, and a hyphen defines an inclusive ascending range. The parser treats each piece of your input in the order it appears, so the order you type is the order the new PDF will follow.
| You type | Pages placed into the new PDF (in order) |
|---|---|
| 5 | 5 |
| 1-3 | 1, 2, 3 |
| 5, 1-3 | 5, 1, 2, 3 |
| 3, 1-3 | 3, 1, 2 (the second 3 is ignored) |
| 5-3 | Rejected — descending ranges are not allowed |
| 2.5, -1, or 600 | Rejected — decimals, negatives, and out-of-range pages are errors |
Repeats follow a deliberate first-occurrence rule. If you list 3, 1-3, the new PDF contains page 3, then page 1, then page 2; the second occurrence of page 3 is ignored, and the result tells you which page number was duplicated. This prevents a casual overlapping range from quietly inflating the output while preserving the order in which each page first appeared. The output page count always equals the number of unique selected pages. If you genuinely need the same page more than once in the new file, use a page-reordering workflow designed for duplication rather than extraction, since the extractor is intentionally one-of-each.
What the New File Keeps and What It Loses
The extractor copies the actual page objects rather than rasterizing them into screenshots. Each selected page keeps its 90-degree rotation, crop box, media box, page graphics, and ordinary page resources such as images and fonts. The page count of the new PDF equals the number of unique pages you selected, and an integration test verifies that the dimensions and identity of each output page match the source it came from. This catches implementations that merely return the right number of pages while copying the wrong pages or sorting your order behind your back.
Some PDF features depend on document-wide relationships that do not survive extraction cleanly. The new PDF is a separate document, so anything that pointed back to the original structure may lose its anchor. The table below summarizes what to expect.
| Feature in the source PDF | Behavior after extraction |
|---|---|
| Page rotation, crop box, media box | Copied with each selected page |
| Embedded fonts and images | Copied as page resources |
| AcroForm fields, annotations, scripts, embedded files | May depend on objects outside the selected pages |
| Bookmarks, page labels, outlines, cross-page links | May depend on document-wide structure |
| Digital signatures | Will not remain valid — the new PDF is a different document |
| Dynamic XFA forms | Not a supported preservation target in pdf-lib |
Before sharing or archiving the result, open it in your target PDF viewer, confirm the page order and appearance, and test any links or fields you rely on. Keep the original PDF in case you need to redo the extraction with different settings or restore a feature that did not survive. If preservation of signatures, XFA, accessibility structure, portfolios, attachments, or regulated archival conformance matters, use a specialist PDF editor designed for that purpose rather than an extractor.
Beginner Mistakes and Quick Fixes
Most beginner errors fall into a few patterns. Treating page numbers as zero-based is the most common one — page 0 does not exist, and typing 0 produces an explicit error rather than a silent skip. Typing a descending range such as 5-3 is the next most common slip; the parser rejects it instead of silently swapping the endpoints, so the typo shows up immediately in the error message rather than producing a surprising output order. Forgetting to separate tokens is another classic mistake: 1, 3-5, 8 works, but reading your own input back without a clear separator can lead to off-by-one guesses about what the parser actually saw.
Another easy mistake is assuming the new PDF will keep document-wide features like bookmarks, digital signatures, or form scripts. Because extraction creates a new file, signatures cannot remain valid, and anything that referenced other pages or attached files may lose its anchor. If your goal is to keep a signed contract intact, use a specialist PDF editor instead of an extractor.
Finally, if your PDF is larger than 50 MiB or has more than 500 pages, the tool will reject it rather than trim it down. Damaged or unsupported files can also fail to load, which is worth keeping in mind for older scans. When the limit is the only issue, you can compress the PDF first so it fits the size boundary, then come back to Extract PDF Pages for the actual page selection. If the problem is page count rather than bytes, splitting the source into smaller documents and extracting from each part gives you the same outcome in two passes.
For a deeper look, see Extract Links From a PDF Explained: What the Tool Reads.
For a deeper look, see PDF Page Counter for Beginners: A Plain-English Guide.