Under the hood
How it works — and where its limits are
Splitting is the mirror image of merging: instead of copying pages from many files into one, this tool
copies selected pages from one file into new documents, using the same pdf-lib copyPages mechanism.
Because pages are copied rather than re-rendered, the extracted pages are identical to the originals — same
text layer, same images, same page size. Nothing is rasterised and nothing is re-encoded, so an extraction is
lossless.
Range mode. The range field accepts a simple, explicit grammar: single page numbers and
hyphenated ranges, separated by commas. 1-3,5,8-10 means pages 1, 2, 3, 5, 8, 9 and 10, in that order,
in one output PDF. Pages are counted from 1, the way readers display them, not from 0 the way programmers count.
Out-of-range numbers are rejected with a message rather than quietly ignored, because a silently shortened
extraction is exactly the kind of error that ends with the wrong pages sent to the wrong person. The tool shows
the document’s total page count next to the file name so you can check your range against it before you run.
Split-every-page mode. Sometimes the job is the opposite: a bundle scan arrives as one long
PDF and each page needs to become its own file — one invoice per file, one certificate per file. This mode
builds a one-page PDF for every page and packs them into a single ZIP download, with zero-padded page numbers in
the file names so they sort correctly in a file manager. The packing happens locally with JSZip; the ZIP is
assembled in your browser and downloaded directly.
What to watch for. As with merging, interactive form fields and the document outline are
not guaranteed to survive page copying with pdf-lib — visible page content always does. Page labels are another
subtlety: if a document uses roman numerals for front matter, the numbers this tool uses are still the physical
page positions (1 is the first page of the file), so count from the start of the file, not from the printed
folio. Encrypted files must be unlocked in a reader first.
Why split locally. The pages people extract are often the sensitive ones — the signed page
of a contract, one statement from a bundle. Here the file never leaves your device: there is no upload, no
server-side copy to expire, and no record of your file’s name or size kept by this site. Keep your original
until you have opened the extraction and confirmed you captured the pages you meant to.