Answer: Load a PDF, pick Ranges (like 1-3, 5, 8-10), Every N pages, or Burst, and each part downloads as its own PDF named after your file. Pages are copied without recompression, but the parts together usually take more space than the original because each one repeats the fonts it uses.
Page ranges, fixed chunks, or one file per page — entirely in your browser
1-3, 5, 8-10 — each range becomes its own PDF), Every N pages (fixed chunks: a 12-page file every 5 pages gives 3 files of pages 1–5, 6–10, 11–12), or Burst (one PDF per page). Pages are copied byte-for-byte, so nothing is recompressed or watermarked, and the file never leaves your device — the split runs locally with the same pdf-lib engine as our merger. Range rules: comma-separated, N-M inclusive, a bare number is a single page, and 9- means page 9 through the end.Click or drag a PDF here
| You type | On a 12-page PDF | Output |
|---|---|---|
1-3 | Pages 1, 2, 3 | 1 file, 3 pages |
1-3, 5 | Pages 1–3, then page 5 | 2 files (3 + 1 pages) |
1-3, 5, 8-10 | Three selections | 3 files (3 + 1 + 3 pages) |
9- | Page 9 through the end | 1 file, 4 pages (9–12) |
-4 (start at 1) | Pages 1 through 4 | 1 file, 4 pages |
| Every 5 pages | Fixed chunks | 3 files (1–5, 6–10, 11–12) |
| Burst | Every page alone | 12 files, 1 page each |
Ranges are inclusive on both ends and may not exceed the document's page count. Command-line equivalents: pdftk in.pdf cat 1-3 output part1.pdf, qpdf in.pdf --pages . 1-3 -- part1.pdf, burst via pdftk in.pdf burst.
Splitting is the mirror of merging: instead of copying many documents' pages into one file, the splitter copies one document's pages into several new ones. Each output file is a valid, standalone PDF carrying its own copy of the fonts, images, and resources its pages reference, which is why the parts together usually take up more disk space than the original, sometimes many times more, even though no content is recompressed (measured below). Page dimensions, orientation, selectable text, and image resolution all pass through untouched.
Take a 12-page quarterly report that needs to go out as three attachments. Loading it shows 12 pages; in Every-5-pages mode the plan table lists three outputs: pages 1–5 (5 pages), 6–10 (5 pages), and 11–12 (2 pages) — 5 + 5 + 2 = 12, every page accounted for exactly once. The files are named after the original, so report.pdf downloads as report-1.pdf, report-2.pdf and report-3.pdf (change the prefix to q1-report first and you get q1-report-1.pdf and so on). Doing the same job in a terminal: qpdf --empty --pages report.pdf 1-5 -- part-1.pdf repeated per range, or pdftk report.pdf burst output page_%02d.pdf for the one-per-page variant.
Splitting and merging are a pair: cut out what matters with ranges, then reassemble in any order — including interleaving pages from different documents — with the PDF Merge tool. That two-step (split → merge) also solves "reorder pages inside one document": split it into ranges in the new order, merge them back.
Bigger than you'd expect, added up. Every part has to carry its own copy of the fonts and images its pages use, and a font used on every page gets copied into every part. We ran this page's own split code (pdf-lib 2.9.1) over 25 real PDFs from our files, cutting each one in half and also bursting it into single pages. Here's a sample, with each total shown as a percentage of the original size:
| Document | Original | Two halves, combined | Burst, all pages combined |
|---|---|---|---|
| 12-page HTML-exported spec | 151 KB | 176% | 942% |
| 47-page research paper | 3.0 MB | 104% | 242% |
| 148-page brand guide | 4.7 MB | 105% | 172% |
| 12-page product brochure | 2.9 MB | 100% | 105% |
| 66-page text playbook | 419 KB | 170% | 4,547% (about 19 MB) |
| 206-page Word-exported report | 2.5 MB | 159% | 1,183% |
Across all 25 files, two halves came to a median of 154% of the original, and a full burst came to 621%. The type of file matters most. Image-heavy files barely grew, because each picture appears on only one page and goes into only one part (median burst 153% across 7 files over 2 MB). Text-heavy files exploded, because they're mostly font data and every single-page file needs those fonts again (median 682% across 16 files under 300 KB).
That matters if you're splitting to get under an email or upload limit. Each part will be smaller than the original, but don't assume a 20 MB file split in two gives you two 10 MB files. Check the sizes in the download list, which shows each file's KB. If a part is still too big, run it through the PDF compressor, and our guide to email attachment size limits lists what each provider accepts.
| Part of the PDF | In the split files | Why |
|---|---|---|
| Text, images, page size | Kept exactly | Pages are copied, not redrawn |
| Links on the page | Kept | They belong to the page |
| Bookmarks (outline) | Dropped | Each part is a brand-new document |
| Fillable form fields | Usually stop working | The form definition lives at document level |
| Title, author, dates | Not copied | New documents start with fresh metadata |
| Password protection | Refused at load | Unlock first, then split |
If you need the bookmarks or a working form in the result, splitting isn't the right tool. Use Delete PDF Pages to remove what you don't want instead, or keep the full file and send a page range note. To pull a handful of non-adjacent pages into one file, Extract PDF Pages does it in one step.
Load the PDF, choose Ranges mode, and type comma-separated ranges like 1-3, 5, 8-10. Each range becomes its own PDF, so a 10-page document with 1-3, 5, 8-10 produces three files with 3, 1, and 3 pages. Open-ended ranges work too: 9- means page 9 to the end.
Choose 'Every N pages' and set the chunk size. The tool cuts at fixed boundaries: a 12-page PDF split every 5 pages yields files containing pages 1-5, 6-10, and 11-12 — three files of 5, 5, and 2 pages. The last file holds whatever remains. This is the mode for print batches, chapter-sized chunks, and upload limits.
Burst mode. One page in, one file out: a 6-page document becomes 6 single-page PDFs, named after your file by default, so report.pdf gives report-1.pdf through report-6.pdf. The command-line equivalent is pdftk in.pdf burst or qpdf --split-pages in.pdf.
No. Pages are copied into new documents byte-for-byte, so text, images, fonts, and page dimensions are untouched. Each output file carries its own copy of the shared resources (fonts, images) that its pages use, so the parts together are usually bigger than the original, but nothing is recompressed or degraded.
Not while it is encrypted. Open the file with its password in a viewer and re-save an unlocked copy (print-to-PDF also works), then split that. The tool tells you when a file cannot be read rather than failing silently.
Splitting turns ranges into multiple output files; extracting pulls selected pages into a single new document. This tool does the first. For the second, split off the pages you want, then combine them — the PDF Merge tool on this site reassembles pieces in any order. The CLI equivalent is qpdf in.pdf --pages . 2,5-7 -- extracted.pdf.
Each part carries its own copy of the fonts and images its pages use. In our test of 25 PDFs, two halves added up to a median 154% of the original, and bursting into single pages came to 621%. Text-heavy files grow the most, because fonts get repeated in every part.
Links on the pages come along. Bookmarks don't, because each part is a new document, and fillable form fields usually stop working. Title, author and dates aren't copied either.