v0.10.0How to use
- Select one PDF file. The total number of pages is shown.
- Enter the pages to extract. Example: 1-3, 5, 8- (separate with commas, a hyphen means a range, and 8- means page 8 through the end).
- The number of selected pages is shown right below the input. If there is an error, the reason appears.
- Click [Extract] and download the new PDF.
Page range examples
| Input | Result |
|---|---|
3 | Page 3 only |
1-5 | Pages 1 to 5 |
1-3, 7 | Pages 1, 2, 3 and 7 |
8- | From page 8 to the last page |
-4 | From the first page to page 4 |
5, 1-2 | In the order 5, 1, 2 |
Example
If you only need the signature pages (48-50) from a 50-page contract, enter "48-50" and save them as a new PDF. If some pages were scanned upside down, you can fix them with Rotate PDF Pages, and to join several files into one, use Merge PDF.
Supported formats, limits and privacy
- Input: one PDF (up to 100 MB). PDFs that are password-protected, have editing restrictions, or are damaged are rejected with a reason when you add them.
- Pages go into the result in the order you enter them. For example, "5, 1-2" gives page 5, then page 1, then page 2, and a duplicated number is used only once.
- If a range goes beyond the total number of pages, the start is greater than the end, or the format is wrong, nothing runs and the reason is shown.
- Text, images and shapes on the pages are copied as they are. Annotations that were visible on screen, such as form field values and stamps, remain only as the appearance saved in the PDF, as part of the page content. Interactive form features, notes, attached files, bookmarks, links within the document and document information (title, author, etc.) are removed from the result (links to external web addresses are kept). Form values that have no saved appearance are not kept. However, when the document specifies that the "form appearance should be redrawn", a viewer redraws it from the entered value rather than from the saved appearance. So a single-line value of English letters and digits (when the font is confirmed to draw those characters as they are) gets a new appearance generated from the entered value. If there are characters other than English letters and digits (such as Korean or Japanese), multi-line values, password fields, choices that have a separate display name, or fonts that cannot be verified, no result is created and the reason is shown, so that no wrong value is included. Keeping such connections could let content from pages you did not select end up in the result file, so they are removed on purpose. Resources such as images and fonts that the selected pages do not use are also left out of the result. Digital signatures are not preserved.
- PDFs rejected for safety: PDFs whose unusual structure means we cannot guarantee that content from unselected pages stays out; pages that use hidden layers (optional content) or whose layer visibility cannot be determined (this tool does not keep layer information, so hidden content could become visible); PDFs whose decompressed size, number of objects or number of resources exceeds what this device can process safely; PDFs whose content is damaged so it cannot be checked to the end; PDFs whose processing takes longer than 10 seconds; PDFs whose form values cannot be carried over exactly; and PDFs where a field marked "Invisible" has a value or appearance, so that whether it is visible depends on the viewer. When a PDF is rejected, no result is created and the reason is shown.
- Limitations: This tool merges, extracts and rotates the pages of the PDFs you select. It does not guarantee the complete removal of hidden information or information you did not select (sanitizing confidential information). The font subsets and character mappings (ToUnicode) used by the pages may contain the character set of the whole document, and the glyphs of fonts that store character shapes as drawings (Type3), the shared symbol dictionary of scanned documents (JBIG2), and the names of layers used by the result pages may also remain. A crafted PDF that uses non-standard references to pull in images or document information from other pages may not be filtered out completely. For non-public documents, check the result yourself before sharing it.
- All processing happens inside your browser, and files are not uploaded to a server. The PDF processor (qpdf, WebAssembly) is downloaded the first time you add a file in a PDF tool, and the browser cache is used after that.
Frequently asked questions
Can I extract several sections at once?
Yes. Join them with commas, as in "1-3, 7, 10-12", and they all go into one file.
How do I save everything except certain pages?
Enter the sections that exclude the pages you want to remove. To remove only page 5 from a 20-page document, enter "1-4, 6-20" or "1-4, 6-".
Can I split a PDF into a separate file for every page?
Currently the tool only makes the selected pages into one PDF. Splitting into individual files may be considered for a later release.