Split a PDF by page, by batch, or by range
Initializing in your browser…
A 60-page report needs section 3, pages 18 to 27, sent on its own as one document.
Input
report.pdf 60 pages · Extract · 18-27
Output
report_pages_18-27.pdf, one 10 page document, text still selectable
Extract makes a single document from everything you typed; Ranges would have made one file per range instead, and the plan panel names the files before anything runs so the two can never be confused. Pages are copied as objects rather than re-rendered, so the type stays sharp and each page keeps its own size and rotation.
Splitting means different things to different people, so this tool asks which one you mean and then shows you the exact list of files it is about to write before it writes any of them. Every page gives one file per page. Every N pages gives fixed size batches, which is what a 200 page scan needs when it has to go out in twenty parts. Ranges gives one file per range you type, so "1-5, 8, 10-12" is three files of five, one and three pages, not nine files of one. Extract gives a single file holding exactly the pages you chose, in the order you chose them, which is the mode for pulling one chapter out of a report. The difference between the last two matters and used to be invisible. A person who types a range wants the range as a document. The plan panel names every output file and the pages it will hold, so what you see on the page and what lands in the ZIP are the same list. Everything you type is accounted for. A range that runs past the end of the document is trimmed and says so; a range written backwards is read backwards and says so; something that is not a page number is refused by name rather than silently rounded down. And when a selection leaves pages out of every output file, the tool says which pages those are.
Use Extract with "18-27" to get those ten pages as a single document, which is what a colleague expecting a section wants rather than ten attachments.
Use Every N pages with a batch size of 20 to turn a 200 page scan into ten files small enough for a size-capped upload.
A scan holding four invoices at pages 1-3, 4-5, 6-9 and 10-12 splits into exactly those four documents with one line typed.
Break a batch of scanned forms into individual pages for filing or categorisation.
The splitter uses pdf-lib. The source document is loaded once, and each output is a new PDFDocument that copyPages brings the chosen pages into, in the order the plan gives, then addPage and save. The result is a valid standalone PDF that does not depend on the source file being present. Because pages are copied as objects rather than re-rendered, nothing is rasterised and nothing is re-encoded: the text stays selectable, the vectors stay sharp, and each page keeps its own size and its own /Rotate entry. A document of mixed Letter, A4, landscape and A5 pages splits into files of exactly those sizes, and a page that arrived rotated leaves rotated.
Output files are named after what they hold: a single page becomes "_page_7", a run becomes "_pages_2-4", and a scattered selection becomes "_part_01". Names are made unique before anything is written, because two members with the same name inside a ZIP means one of them is lost.
Document-level structures are not carried into the parts. A PDF keeps its bookmarks and its form fields in the catalog rather than on the pages, so copying pages leaves them behind; the same is true of any digital signature, which covers the hash of the whole original document and cannot survive being taken apart.
Downloads are bundled as a ZIP by default, because most browsers block a page that starts several downloads at once. Individual files can be downloaded one at a time from the plan list after the split has run.
Ranges makes one file per range you type: "1-5, 8" gives two files. Extract makes one file containing everything you typed: "1-5, 8" gives a single six page document. The plan panel shows which you are about to get, so there is no need to guess.
No. Pages are copied as PDF objects with no re-encoding, so text stays selectable and vector graphics stay sharp at any zoom. Each page also keeps its own paper size and its own rotation.
You are told. A range past the end of the document is trimmed to fit and the trim is reported; a backwards range like 4-2 is read as pages 4 down to 2 and says so; a word or a decimal is refused by name rather than being rounded to a page you did not ask for. Nothing is silently discarded.
No. A PDF stores bookmarks and form fields in the document catalog rather than on the pages, so copying pages into a new file cannot bring them along. The same applies to digital signatures, which cover the whole original document.
Because most browsers block a page that starts more than one download at a time, so several separate files would arrive as one file and a blocked-popup warning. After the split has run you can also download any single file from the plan list.
PDF parsing and editing happen in your browser. Documents, and everything inside them, are never uploaded or stored remotely.