ALL PDF TOOLS37
Extract PDF Pages
Extract Pages from a PDF — Free Online Tool
Pick the pages you want and save them as a single PDF — or as one PDF per page, zipped.
About Extract PDF Pages
Extracting pages is two different jobs wearing one name, and this tool does both. Drop a PDF and you get a deck of page thumbnails; click the pages you want, or type them into the RANGE field as something like 1-3, 5, 8-10 — the deck and the field are the same selection in two shapes, so typing updates the tiles and clicking rewrites the text. Select all and Deselect all sit in the rail next to them. Then pick the output mode. One PDF gives you a single document holding the selected pages in their original order, which is the ordinary case. One per page gives you a separate PDF for each page you picked, named after the page number it came from — pick pages 3 and 9 and you get page-3.pdf and page-9.pdf, delivered together as a ZIP. Select exactly one page in that mode and there is nothing to zip, so the file saves on its own instead, and the readout says so before you press anything. The pages themselves are grafted across whole by pdf-lib rather than redrawn: fonts, embedded images and vector art come over exactly as they were, annotations travel with the page they sit on, and nothing is re-encoded or rasterised at any point. What does not come across is the layer that belongs to the document rather than to any page — the outline tree, the interactive form, named destinations and page labels. So an extracted file starts with no bookmarks, and if the source was a fillable form the result still draws every field and whatever was typed into it but is no longer a form: the boxes cannot be filled in or read back by a form tool. Front matter numbered i, ii, iii comes back as 1, 2, 3 for the same reason. Internal cross-references are followed through rather than left to fend for themselves: a link whose target page is one of the ones you picked is repointed at it and still works, and a link to a page you left behind has its destination stripped, so it does nothing instead of quietly jumping to the front of the document. A link that named its target rather than pointing straight at it goes dead either way, because the table of names is part of the same document layer that does not travel. All of that is equally true of merging and splitting here. Everything runs in your browser; the file never leaves your machine.
Questions
How is this different from Split PDF?
Split PDF does one thing: the pages you pick come out as a single new document. This tool does that too, and adds the half it does not have — One per page mode writes a separate PDF for every page you selected and hands them over as one ZIP. The page deck, the range syntax and the selection behaviour are identical between the two.
How are the separate PDFs named?
After the page they came from, not their position in the batch. Select pages 3 and 9 and the ZIP holds page-3.pdf and page-9.pdf — which is the only naming that still makes sense once the files are unzipped next to the original document.
What happens if I select just one page in One per page mode?
There is nothing to zip, so the single PDF saves on its own as page-N.pdf rather than inside an archive. The output readout shows that exact filename before you run anything, so the button never promises a ZIP it will not produce.
Do the pages come out in the order I clicked them?
No. The selection is sorted before extraction, so pages always keep the order they had in the original document — clicking 5, then 1, then 3 gives you 1, 3, 5. To change the sequence itself, run the result through Reorder PDF Pages.
Are bookmarks and formatting preserved?
Formatting is: pages are grafted across whole, so fonts, embedded images and vector art come over untouched, each page's annotations travel with it, and nothing is re-encoded. What is not preserved is anything that belongs to the document rather than to a page — the outline tree, the interactive form, named destinations and page labels. So an extracted file starts with no bookmarks, and a fillable form still draws its fields and whatever was typed into them but can no longer be filled in or read by a form tool. Links between the pages you extracted keep working: each one is repointed at the page it names. A link to a page you did not extract has its destination removed, so it does nothing rather than jumping to the front of the file, and a link that pointed at a named destination goes dead because the table of names is part of the document layer that does not travel. The same is true of merging and splitting here.
Is my file uploaded to a server?
No. Transmute processes everything locally in your browser using JavaScript and WebAssembly. Your files never leave your device — there is no server, no upload, no cloud processing.