Split PDF Without Uploading

Splitting takes one document and produces several, and everything happens inside your browser: the file is parsed in memory with pdf-lib, the pages you selected are copied into new documents, and each result is offered as a download. Nothing is uploaded, which is why this tool is safe for the documents people actually need to split: signed contracts where one exhibit must go to one party, scanned bundles from an administration, or a year of statements that should become twelve files.

There are two ways to think about a split. Extraction pulls a chosen range into a new file and leaves the rest behind: pages 12 to 18 become their own document. Separation cuts the whole document at boundaries: every page, or every N pages, becomes a file. Both produce genuine PDFs with selectable text, not images of pages, because pages are copied as live objects rather than re-rendered.

The pages you extract keep their internal resources: fonts, images and vector drawings travel with the page. That means extracting one page from a 200-page report can produce a file larger than one two-hundredth of the original, since shared fonts must be embedded in each output. It is normal, not a bug, and it is the price of every output file standing alone.

What splitting cannot rescue: page labels and bookmarks. If the original numbers its pages 'iv, v, 12-40', the outputs count from 1 again in the viewer's sidebar, and outlines pointing at removed pages disappear. Check the result before sending it to a court or a registry that cares about pagination.

Memory is the practical ceiling. A split reads the entire source into RAM, so very large scanned files, hundreds of megabytes of full-page images, can strain a phone. On a laptop, splitting even long documents is quick; the slowest part is usually your own decision about where to cut.

If the file asks for a password when opened, unlock it first with the Unlock PDF tool; a reader that cannot open the file cannot copy its pages either.