split PDFPDF splitterpage rangesPDF tools

How to Split a PDF by Pages, Ranges, or File Size

PDF splitting can export every page, split selected page ranges, or create smaller parts to satisfy upload size limits.

ToolGarden tools prioritize browser-local processing, so files and text do not need to be uploaded to a server.

Published July 2, 2026Updated July 3, 20268 min readBy ToolGarden

Before splitting a PDF, decide whether you need every page as a file, several page ranges, or smaller parts for an upload limit.

Three Common Split Modes

Mode	Example	Best for
Every page	Export pages 1, 2, 3 separately	Submitting scanned pages one by one
Page ranges	1-3, 4-8, 9-12	Splitting by chapters or attachments
By size target	Keep each part below a limit	5MB or 10MB upload limits

Exact automatic size-based splitting is hard because scanned pages, photos, and text pages vary widely. A practical workflow is to split by ranges, then check each output file size.

Why Is Size-Based Splitting Not Exact?

PDF page size can vary dramatically. A text-only page may be only a few dozen KB, while a scanned image page can be several MB. Even with the same page count, output files can have very different sizes.

Page type	Size behavior	Splitting tip
Text-only PDF	Usually small and stable	Split by chapter ranges
Scanned PDF	One page can be large	Test with a small page range first
Image-heavy report	Chart pages are larger	Separate image-heavy sections
Mixed document	Each page can differ greatly	Check each output file size after splitting

How to Write Page Ranges

A common syntax is 1-3,5,8-10, which means pages 1 through 3, page 5, and pages 8 through 10. Check whether PDF page numbers match the reader display before exporting.

1-3 means a continuous range and includes both page 1 and page 3.
1,3,5 selects only those pages.
1-3,8-10 combines multiple ranges in the output.
If the viewer shows Roman numerals for front matter, displayed page labels may not match actual PDF pages.

How Should You Name Split Files?

If split files will be uploaded or sent to someone else, use names that preserve meaning and order. contract-part-01.pdf and contract-part-02.pdf are much safer than split-1.pdf.

When Should You Extract Pages Instead?

If you only need pages 2, 5, and 9 in one new document, extraction is a better fit. If you need one large PDF broken into multiple independent files, splitting is the right workflow.

Frequently asked questions

Q.Why are my split PDFs each larger than expected?

Two reasons dominate. First, the source PDF likely embeds whole fonts and shared images once; when you split, each output has to embed its own copy of the resources it references. Second, most splitters don't re-run compression or object deduplication on the outputs, so the writer leaves some overhead. If the total across splits is significantly bigger than the source, run each split through a PDF compressor afterwards. It's also worth checking whether contiguous pages could share resources; splitting into fewer, larger files is usually smaller in aggregate than many tiny ones.

Q.Does splitting by file size hit the target exactly?

No. PDF pages share fonts, images, and other objects, and each page's contribution to the file depends on how much is unique to it. Splitters accumulate pages until adding the next page would exceed the target, then cut. That means individual outputs can be either slightly under the target or slightly over the previous cut, depending on the boundary page. If you need a hard cap (say, an email attachment limit), set the target 15-20% below the real limit to give yourself a safety buffer.

Q.Why do bookmarks disappear when I split a PDF?

PDF bookmarks (outlines) are a tree of pointers into the document by page. When you split, a bookmark that pointed to page 20 no longer makes sense in a 10-page slice, so most tools drop the outline entirely rather than fix up destinations. Preserving a localized outline requires the tool to walk the tree and rewrite each entry to reference only the pages that ended up in each output; pdf-lib does not do this automatically. If bookmarks matter, keep the original whole and use ranges/extract-pages instead.

Q.What page-range syntax works reliably across splitters?

Stick to ASCII `,` and `-`; e.g., `1-3,7,12-14`. Avoid full-width Chinese punctuation (`，`, `: `), em-dashes, and en-dashes; those get rejected as invalid characters. Ranges must be increasing (`5-3` fails), and endpoints must be within the document (`1-999` on a 20-page PDF is invalid). Open-ended forms like `10-` (meaning "page 10 to end") aren't supported everywhere; always write full ranges. And remember that page numbering starts at 1, not 0.

Q.Should I split-then-merge, or use extract-pages, to pull selected pages?

Use extract-pages. Splitting produces multiple files; if what you actually want is a single new PDF containing, say, pages 1-3 + 7 + 12-14, splitting forces you to then merge those splits back together, which duplicates work and often loses metadata twice. Dedicated extract-pages tools accept comma-separated ranges and produce one PDF with just those pages, preserving order. On ToolGarden that's `/pdf/extract-pages`; use split only when you truly need N separate output files.