How to Split a PDF: Extract Pages, Ranges or Every Page Separately
Splitting, extracting and deleting pages are three different operations. Which one you want, how to get the page numbers right, and what does not survive the split.
Updated August 11, 20265 min read
"Split this PDF" describes at least three different jobs: pull one section out into its own file, chop a document into many single-page files, or simply delete the pages you do not want. They produce completely different results and it is worth being clear which you mean before you start — most of the frustration with page tools comes from running the right operation on the wrong mental model.
Three operations, three outcomes
| Operation | Input | Output | Use when |
|---|---|---|---|
| Extract | 1 file | 1 new file with the pages you chose | Sending only the appendix |
| Split into parts | 1 file | Several files | Breaking a scan into per-chapter documents |
| Remove | 1 file | 1 file with pages deleted | Dropping blank or duplicate pages |
The distinction that matters most: extract keeps what you select and discards the rest, while remove discards what you select and keeps the rest. Running the wrong one gives you the exact inverse of what you wanted, which is easy to miss if you do not open the result.
Getting the page numbers right
This is where most split jobs go wrong, and the cause is nearly always the same: the page numbers printed on the page are not the page numbers the tool uses. A report with a cover, a contents page and two blank sheets before "page 1" means printed page 1 is document page 5.
Tools count physical sheets from 1, ignoring whatever is printed on them. Roman-numeral front matter, unnumbered dividers and scanned inserts all shift the offset. Before entering a range, scroll to the first page you want and read the *viewer's* page counter, not the page itself.
The same trap catches people going the other way. If a colleague asks for "pages 40 to 55" from a printed copy, confirm which numbering they mean before extracting — being off by four pages in a legal or financial document is a real problem, not a cosmetic one.
The workflow
Fix the page order first if it is wrong
Splitting a document whose pages are out of sequence just gives you several files that are also out of sequence. Sort it once with the page reorder tool before splitting.
Note the real page numbers
Use the viewer's counter. Write down the start and end of each range you need before you open the tool.
Extract the range
In the PDF splitter, enter the range and produce a new file. For several sections, choose Page ranges and type them all at once — 1-5, 10-12, 40-55 gives one file per range — or tick the pages you want and they are gathered into a single file.
Open every output
Check the first and last page of each file. An off-by-one at a boundary is the single most common error and it is invisible unless you look.
Runs entirely on your device — contracts, statements and medical records are never uploaded.
Split a PDFWhat does not survive
Page content copies across intact: text stays as text, images keep their resolution, and nothing is re-compressed. What can be lost lives at the document level rather than on the page:
- Bookmarks — our splitter copies none of them into the new files, not even those pointing at pages the part still contains.
- Internal links — a cross-reference to page 60 in a file that now ends at page 20 has nowhere to go.
- Digital signatures, which always break: a signature attests to one specific document and an extracted subset is not that document.
- Form field relationships, where fields calculated from values on other pages lose their inputs.
One thing that emphatically does survive is metadata. The extracted file inherits the original's title, author, creating software and creation date — so a three-page extract from a confidential report still carries the report's identity. Check it with the metadata editor before sending anything outside your organisation.
Splitting is not redaction
Removing pages genuinely removes their content — that part is real. But it is a page-level instrument, and people reach for it when the sensitive material is a paragraph rather than a page. Extracting pages 1–5 to avoid sending page 6 works; blacking out a name on page 3 and then splitting does not, because the text under the black box travels with the page.
If the goal is to remove specific content rather than whole pages, that is proper redaction, which deletes the underlying objects instead of drawing over them.
And if the real problem is that the file is too large rather than too long, splitting only helps when the recipient genuinely needs a subset. Otherwise compressing it keeps the document whole — scanned documents routinely give up 80–95% with the text still crisp.
Finally, name the outputs properly as you go. A folder of document (1).pdf through document (9).pdf is indistinguishable a week later, and page ranges are exactly the kind of detail nobody remembers. contract-appendix-b-p40-55.pdf costs a few seconds and tells you what it is, where it came from and which pages it holds without opening it.
Frequently asked questions
How do I extract specific pages from a PDF?
Open the file in a splitter, enter the page range you want, and export it as a new document. Take the numbers from your viewer's page counter rather than the numbers printed on the pages — cover sheets and front matter usually make the two differ.
What is the difference between splitting and removing pages?
Extracting keeps the pages you select and discards the rest, producing a new smaller document. Removing deletes the pages you select and keeps everything else in one file. They are inverses, so picking the wrong one gives you exactly the opposite of what you wanted.
Does splitting a PDF reduce quality?
No. Page content is copied across unchanged — text stays vector, images keep their original resolution, and nothing is re-compressed. The output is smaller only because it contains fewer pages.
Is it safe to split confidential PDFs online?
Only if the tool does not upload them. Ours processes the file on your own device. Do check the metadata of the extract before sending it on — title, author and creation history are inherited from the original document.