Split PDF

Extract pages or cut a PDF into several files. Free, no upload. 1 GB max · Up to 1 GB · Processed locally, never sent to a server.

On a sixty-page lease, only pages 12 to 14 interest your insurer. On a batch of payslips scanned in a single pass, every page has to become a separate file instead. Split PDF covers both needs with two distinct modes, and an input field whose syntax is worth knowing before you start the run.

The mechanism is the same in both cases: the document is opened with pdf-lib, a new document is created, and the requested pages are copied into it. That copy is faithful. There is no rasterisation, no re-encoding and no recompression: text stays selectable, images keep their original resolution, and nothing is stamped onto the pages produced.

The split happens on your own device, which counts when you only want to pass on three pages of a medical file or an internal report without handing the other fifty-seven to a third-party service. The history shown under the tool keeps nothing but the name and the size of the last four results, in the browser’s local storage, never their content.

How to use it

  1. Drop in the PDF to cut One document at a time. If it is encrypted, reading fails at the opening stage with a message about a damaged or password-protected PDF.
  2. Choose the mode “Extract a page range” produces a single file. “One page per file (ZIP)” produces as many documents as there are pages, and hides the input field.
  3. Type the pages you want The Pages field contains 1 by default. It accepts single numbers and intervals separated by commas, on the model 1-3, 5, 8-10.
  4. Click Run In range mode the progress line announces how many pages were kept. In ZIP mode it counts through page by page to the last one in the document.
  5. Download the file or files A single file carries a note along the lines of “3 page(s) of 60”. A series of files comes with the “↓ Download all (ZIP)” button, on top of the individual downloads.

The exact syntax of the Pages field

The separator is the comma, and spaces around the values are tolerated. An interval is written with a hyphen between two numbers; typed backwards, it is corrected silently, so 10-4 is the same as 4-10. Numbers that fall outside the document are dropped without warning, and duplicates are removed. Leaving the field empty does not mean all pages: the default value, 1, takes over instead.

Two error messages can appear. The first, about an invalid page range, fires on anything that is neither a number nor an interval: a semicolon, the word “to”, or an em dash pasted in from a word processor, since only the hyphen from the keyboard is recognised. The second says that no valid page is left after filtering, and reminds you how many pages the document has.

Output order and file names

The selection is sorted in ascending order and deduplicated before extraction: an entry such as 5, 1, 1 gives page 1 then page 5. This tool cannot reverse pages, and it cannot duplicate one. To obtain a different order, extract each block separately, then assemble the files with Merge PDF.

The name of the file produced repeats the name of the document, followed by -pages- and then the first and last page kept. An entry of 1, 9 therefore produces a name carrying those two bounds even though the file contains only two pages: the note displayed under the result is what counts.

The one-page-per-file mode

Each page becomes a standalone document, named after the starting file and suffixed with its number on two digits, with a leading zero up to page 9. The files are listed one by one with their own download button, and an extra button groups them into a ZIP archive assembled by JSZip with DEFLATE compression.

That mode has a cost: the documents are built one after another and held in memory until they are downloaded. On several hundred pages, and particularly on a phone, range mode stays far cheaper.

What splitting does not do

It does not lighten things proportionally. Each file produced re-embeds the fonts and the images its own page depends on, so the parts added together weigh more than the whole. To bring their size down, send them through Compress PDF.

It does not retouch the pages it keeps either. A page saved in landscape comes out in landscape, a rotated page keeps its rotation, and nothing is recalculated. If your aim was to straighten scans that came out crooked, Rotate PDF is the tool to use: its Pages field takes the same syntax, except that an empty field there really does mean the whole document.

Frequently asked questions

How do I extract a single page?

Type its number on its own in the field, 7 for instance. The file you get carries that number in its name and contains only that page, at its original size.

What happens if I ask for a page that does not exist?

The out-of-range number is ignored without warning, and the rest of the entry is processed as normal. If the whole entry falls outside the document, the run stops and reminds you how many pages are available.

Can I put the pages back in a different order?

No, the selection is always sorted before extraction. Work in successive blocks, then rebuild the order you want with Merge PDF, which follows the order of its list.

The field rejects the entry 1–3. Why?

Because the dash it expects is the hyphen from the keyboard. An en dash or em dash, often introduced by a word processor’s autocorrect, is not recognised and triggers the invalid range message.

Does the text in the extracted pages stay selectable?

Yes. The pages are copied as they are, with no detour through an image: text, fonts and vector artwork are identical to those in the starting document.

My document runs to three hundred pages. Is ZIP mode sensible?

It will work, but three hundred documents will be built and then held in memory before the archive is assembled. On a modest machine, prefer several passes in range mode.

Similar tools

Discover other tools

Categories