Split a PDF three ways
Take one document apart: every page as its own file, hand-picked ranges such as 1-3,5,7-9, or equal batches of N pages. Each part is generated separately and downloads on its own.
Every page becomes its own PDF, named file-page-1.pdf, file-page-2.pdf and so on. Useful for separating a batch of signed forms or pulling single pages out of a scan.
Each comma-separated group becomes one PDF. 1-3,5,7-9 produces three files; 4 produces one. Spaces and the dashes pasted from Word are handled.
A 32-page document split at 10 pages gives 10, 10, 10 and 2 - the last part is never padded.
Which mode fits your job
All three modes copy pages into new documents; none of them re-renders the page. That means text remains selectable and image quality is identical to the source. The only real decision is how the pages are grouped.
| Job | Mode | Why |
|---|---|---|
| Send page 3 of a contract to a colleague | Page ranges, value 3 | One file, no leftovers to delete |
| Separate 40 scanned forms | One file per page | Each recipient gets exactly one page |
| Email a 300-page manual | Every N pages | Attachments stay under the mail limit |
| Reassemble chapters after a merge | Page ranges | Chapter boundaries rarely fall on a fixed interval |
| Import into a system with a page limit | Every N pages | Predictable part sizes |
Range syntax in detail
The range field takes one or more parts separated by commas, semicolons or new lines. A part is either a single page (7) or a span written low to high (7-9). Everything else is rejected with an explanation rather than silently ignored - a mistyped range that produces the wrong pages is worse than an error message.
| You type | You get |
|---|---|
1-3,5,7-9 | Three files: pages 1-3, page 5, pages 7-9 |
2;4;6 | Three single-page files |
10-8 | An error - spans must run low to high |
1-40 on a 12-page file | An error naming the last page (12) |
1-3, 2-4 | Two files with overlapping pages; that is allowed |
Overlaps are deliberate. If two departments need overlapping extracts, you can produce both in one pass instead of running the tool twice.
File naming you can still sort later
Parts are named from the source file with a zero-padded index where the order matters, for example manual-part-01-pages-1-10.pdf. Zero padding keeps part-09 before part-10 in every file manager and every script that sorts names as text. Source names are lowercased and stripped of spaces, so a file called Q3 Report (final).pdf becomes q3-report-final.
Downloading many files at once
Browsers limit how many downloads a page can start in a row, and some ask for permission the first time. "Download all" spaces the requests out, and each part also has its own button, so you can fetch the important ones first. If a batch stops halfway, click the remaining buttons individually rather than re-running the split.
Limits worth knowing
- A single run is capped at 300 output files. Above that, split in batches with the range mode.
- Every part is built in memory before you see the list, so a 400-page document split per page uses a lot of RAM. On a phone, prefer the range mode.
- Encrypted documents may open in restricted mode. If the output pages are blank, remove the password first.
- Splitting does not reduce total size much: each part carries its own copy of shared fonts and images.
The longer explanation, with the arithmetic for chunk sizes and naming edge cases, is in Splitting a PDF: single pages, ranges and chunks. To go the other way, use the merge tool.
Split FAQ
Can I split a PDF into two halves?
Yes - use page ranges and type both halves, for example 1-6,7-12. You get exactly two files, split wherever you want.
Do the parts keep the original page size?
Yes. Page boxes, rotation and content are copied unchanged, so mixed-size documents keep their mixed sizes.
Why is my range rejected?
The most common reasons are a reversed span (9-7), a page number past the end of the file, or a stray character such as a page label like p.3. Use plain numbers and hyphens only.
Is the original file changed?
Never. The source stays on disk untouched; only new files are created.