What this tool does
Splitting a PDF sounds trivial and is where most tools get quietly wrong, because the input is a little language rather than a number. People write 1-3, people write 1–3 with the dash they copied out of an email, people write -3 to mean the last three pages, and people write page 90 of a twelve-page document without noticing. A tool that guesses instead of checking will hand you the wrong pages.
How it works
The expression is parsed rather than split on a dash, and the parser is strict on purpose. It accepts comma-separated parts such as 1-3,5,8-10, an open-ended part such as 8- meaning from page 8 to the end, and a leading dash such as -3 meaning the final three pages. It tolerates the seven different dash characters a word processor can produce, because 1–3 with an en dash is a different character from 1-3 with a hyphen and people paste the first one. It rejects reversed ranges like 5-2 rather than guessing, since a guess in a tool that removes pages is a guess about what to destroy.
Every parse reports what it did before anything is written. The line under the input says exactly which pages matched, and if the expression asked for a page that does not exist it says which one and that it is being ignored, rather than quietly returning a shorter result. That is the difference between a tool that tells you what it did and one that makes you count the pages of the output to find out.
Two output modes, because two jobs get described as the same thing. Splitting by range writes one file per expression part, so 1-3,7 produces three files and that is what you want for separating chapters or pulling an appendix out of a report. Splitting by page writes one file per individual page, which is what you want for a batch of scans that arrived as a single document. Choosing the wrong one produces thirty files when you wanted one, so both are stated on the control rather than guessed.
The pages are copied rather than re-rendered, so an extracted page keeps its selectable text and its vector graphics. Splitting is also non-destructive in the useful sense: the source file is never modified, and there is nothing to undo because nothing was ever changed.
Worked example
A 200-page report has to go out as the summary, the methodology appendix and the tables, and the tables are the last nine pages.
- Enter 1-6 for the summary and check the preview confirms six pages
- Add the appendix pages and enter 192-200 for the tables, or -9 for the last nine
- The preview reads back the page numbers before anything is written, so a typo is visible immediately
- Split by range to get one file per part, or by page to get every page separately
The parts come out as separate files in one archive, in the order you listed them, with a page count you already confirmed rather than discovered afterwards.
Accuracy and limitations
- An expression that selects no pages produces nothing. The tool says so rather than returning an empty file.
- Pages beyond the end of the document are ignored and reported, so a file can come out shorter than the expression implied.
- Splitting by page on a long document produces one file per page, which is a lot of files. Use the range mode unless you genuinely want every page separately.
- A password-protected PDF cannot be opened here, so splitting one is not possible.
Frequently asked questions
- How do I extract specific pages from a PDF?
- Type the page numbers in the input, using commas and dashes, such as 1-3,5,8-10. The line underneath reads back exactly which pages matched before you do anything, so you can confirm the parse is right rather than counting pages in the output afterwards.
- What does -3 mean in a page range?
- It means the last three pages, counting back from the end. An open-ended range works the other way, so 8- means page 8 through to the final page. The two forms together cover the common jobs without needing to know how long the document is.
- Why does my range not work?
- Reversed ranges are rejected on purpose, because 5-2 is a typo and guessing which way round you meant would risk giving you the wrong pages from a tool whose output is a set of files. Write it as 2-5. A range that runs past the end of the document is accepted but reported, so you can see which page was ignored.
- Should I split by range or by page?
- By range if you want one file per part you named, which is what you need for chapters or appendices. By page if every single page should be its own file, which is what you need for a batch of separate scans that arrived as one document. Splitting by page on a long report produces a lot of files.
- Does splitting reduce quality?
- No. Pages are copied into the new files rather than re-rendered, so text stays selectable, vector graphics stay sharp and a scanned page keeps its original resolution. Nothing about the page is recompressed.
- Are my PDFs uploaded to be split?
- No. The document is read, split and re-packaged in your browser, and the original file on your computer is never modified. Nothing is sent anywhere at any point.