How to Split a PDF Into Separate Files or Pull Out a Single Page

A PDF is a container. Most of the time that's convenient — one file holds the whole contract, the whole scanned book, the whole board pack. But the moment someone asks for just page 38, or only the finance section, or each chapter as its own file, the container works against you. You don't want to email a 62-page signed agreement when the recipient needs one rider. You want to reach in, take what you need, and leave the rest closed.

That reaching-in is what splitting does. This guide covers the three ways to do it — one file per page, by page range, and by file size — and explains why doing the work inside your own browser, instead of uploading to a server, is the part that actually matters for sensitive documents.

The three ways to split, and when each one fits

One PDF per page. Drop a 24-page contract in and get back 24 single-page PDFs. This sounds blunt, but it's exactly what records systems want: one document per page so finance can be granted access to the payment-terms page without seeing the indemnity page. The PDF Splitter names each file with its page number, so contract-p1.pdf through contract-p24.pdf land in a single zip, ready to bulk-import.

By page range. This is the workhorse. You pick the page ranges — something like 1-5, 8, 12-20 — and each comma-separated chunk becomes its own new PDF. Single pages and ranges mix freely. The syntax follows Acrobat convention: 1-indexed (page 1 is the cover), commas or spaces separate items, 5- means "page 5 to the end", and -3 means "page 1 through 3". Type a range wider than the document, like 1-999 on a 10-page file, and the extra pages are quietly dropped rather than erroring out.

By file size. Set a megabyte ceiling and the tool greedily packs pages into chunks, cutting each time the next page would push a chunk over the limit. This is the one you want when an email gateway rejects anything over 10 MB and you just need the whole thing in pieces that will send.

A worked example: pulling pages 3 to 7 out of a report

Say you have quarterly-review.pdf, a 30-page report, and a colleague only needs the methodology section — pages 3 through 7. Here's the whole flow:

Open the PDF Splitter and drop quarterly-review.pdf onto it. The page count appears immediately, so you can confirm it's the file you meant.
Switch the mode to By page range.
Type 3-7 into the range field. One chunk, one output file.
Hit split. You get back quarterly-review-p3-7.pdf — a clean 5-page PDF containing exactly those pages, embedded fonts and images intact, downloaded straight to your machine.

That's it. The five pages come out byte-faithful to the source: copyPages carries the full resource graph for each page, so nothing about how those pages look changes. If you'd wanted methodology and the appendix on pages 28-30 in the same file, you'd type 3-7, 28-30 and the chunk would contain all eight pages stitched in order.

The first time I actually needed this

I'll be honest about where this clicked for me. A lawyer sent me a 62-page master agreement and the client's insurer needed exactly one page — the indemnity rider on page 38. My instinct was to open the file, screenshot the page, and paste it into a document. That would have destroyed the text layer, the signature block, everything that made it a real legal page. The alternative I'd seen people reach for was worse: upload the entire signed contract to whatever free splitter came up first in search.

Instead I dropped the PDF in, typed 38, and had master-agreement-2026-p38.pdf on my desktop in under a second — the page standing alone, identical to what was signed. The signed contract never left my laptop. That second part is the whole point, and it's worth spelling out.

Why splitting locally keeps documents private

Most online PDF tools work by uploading your file to their server, processing it there, and handing back a download link. For a meme image, fine. For a signed contract, a scanned medical record, or a board pack full of unannounced numbers, you've just handed a copy of a confidential document to a third party you know nothing about — their retention policy, their breach history, who can read the bucket.

Local splitting closes that hole entirely. The file is read with File.arrayBuffer() straight into memory, pdf-lib copies the pages you asked for into new documents, an inline zip writer bundles them, and one download fires. The PDF, its text, its embedded images, and any signatures inside it never touch a server — because there isn't a server that could receive them. If you want to confirm it yourself, open DevTools, switch to the Network tab, and run a split. The request count stays at zero.

Privacy note. Splitting happens entirely in your browser. The bytes are read locally, the new PDFs are built locally, and the download is local. Nothing is uploaded. One honest caveat: pdf-lib carries the source document's information dictionary — creator software, modification timestamp, sometimes an author name — onto each output chunk. If that metadata is sensitive, run the output through qpdf --linearize with a metadata cleaner afterward. The split itself adds nothing new.

There's a real trade-off worth knowing. Because everything runs in your tab, the ceiling is your browser's memory: pdf-lib needs the source and the output PDF in RAM at once, so a mid-range laptop handles 200–300 MB of input comfortably, while phones cap closer to 50–100 MB. For an 800-page scanned book, split in two passes — first by range into halves, then by page. That's a small price for never uploading the document.

Putting the pieces back together

Splitting and merging are two halves of the same workflow, and you'll often want both. Carve a board report into per-recipient packets one week, then reassemble approved sections into a final deliverable the next. When you need to recombine — or when filename sort order won't reliably reconstruct the original (p10-12 often sorts before p2-4 in a file manager) — the PDF Merger re-orders by the sequence you specify rather than trusting alphabetical filenames. Pair the two and you can take any PDF apart, keep only what matters, and rebuild it exactly how you want, all without a single upload.

A few habits keep splits clean: remember page 1 is the cover, so 0-3 drops the 0 and gives you pages 1-3 — write 2- to skip the cover. Decrypt user-password PDFs with qpdf --password=… before splitting, since the tool can parse owner-locked files but not truly encrypted content. And if you plan to reassemble by filename later, zero-pad your ranges so natural sort order survives.

Split smart, keep it local, and the container stops working against you.

Made by Toolora · Updated 2026-06-13