PDF workflows guide
PDF Split vs Extract vs Delete: Which Page Tool Should You Use?
The tools look similar, but they produce different outputs. Decide whether you need many files, one selected file, or the original with a few pages removed.
Updated:
PDF page tools are easiest to choose when you start with the output, not the button label. Do you need many separate PDFs, one new PDF containing selected pages, or nearly the same document with a few pages removed? Those are three different jobs.
The quick decision
| What you want at the end | Use |
|---|---|
| One source becomes several files | Split PDF |
| A new file keeps only selected pages | Extract PDF pages |
| One file stays mostly intact but loses a few pages | Delete PDF pages |
| Pages from several source files become one file | Merge PDF |
The same page range can therefore lead to different choices. “Pages 1–3” is an extraction request if it becomes one handout, but a split request if it becomes the first file in a series of chapter files.
Split: one source into multiple output files
Use Split when each part needs to be delivered, uploaded, or archived separately. Common patterns include one file per page, one file per chapter, or ranges such as 1-10, 11-20, 21-30.
Before splitting, decide whether the output names need a stable order. Prefixes such as 01-introduction, 02-method, and 03-appendix prevent a file system from sorting 10 before 2. Afterward, open the first and last page of every output and confirm that no range overlaps or gaps were introduced.
Splitting a document does not automatically reduce the amount of information. If the original contains a 100-page scan and you split it into 100 files, the total bytes can remain almost unchanged. If the receiving service has a total-size limit, extracting only the required pages or compressing the outputs may be a better workflow.
Extract: selected pages into one clean file
Use Extract when the reader needs a curated subset in one PDF. It is the natural choice for a report appendix, a signed page, a chapter, or a form packet assembled from a longer document. Typical input is 1, 5-7, 12.
Check whether the page numbers in the viewer match printed page numbers. A PDF may show a physical page 8 while the footer says “6”. Use the PDF page index shown by the tool, then confirm the first and last visible content in the output. Keep the original because extraction intentionally discards every unselected page from the new copy.
Delete: preserve most of the document and remove a few pages
Use Delete when the source is almost correct. This is useful for blank scans, a duplicate page, a draft cover, or a page containing an accidental personal note. List the pages to remove and verify that the remaining sequence is continuous.
Deletion is not a redaction method. Removing a page does not remove sensitive words that appear elsewhere, inside attachments, in metadata, or in hidden layers. For confidential documents, search the output and inspect metadata before sharing it. If a page contains only part of a sensitive record, use a proper redaction workflow rather than drawing a white rectangle over text.
Worked examples
“Keep pages 3, 8–10, and 22 as one file”
Choose Extract and enter 3, 8-10, 22. The expected output has five pages. Count them: one page + three pages + one page = five. If the output has four or six pages, stop and inspect the range syntax before sending it.
“Remove blank pages 2 and 17 from a 20-page scan”
Choose Delete and remove 2, 17. The expected output has 18 pages. Check the page after each deletion; a blank page may have been intentionally used as a chapter separator.
“Create files for chapters 1–4”
Choose Split with the ranges that match the chapter boundaries. Name each output with a numeric prefix, then open every file. If a chapter starts halfway through a scanned spread, consider extracting and reviewing that boundary manually.
A verification checklist for every output
- Count pages and compare the count with the intended ranges.
- Check the first and last page; these are where off-by-one errors show up.
- Search for a heading that should be present and one that should be absent.
- Confirm orientation, links, annotations, forms, bookmarks, and signatures if they matter.
- Keep the source PDF until the recipient confirms the replacement is usable.
For a private workflow, use browser-local PDF processing where possible and read the tool's limits before selecting a very large scan. The PDF upload troubleshooting guide covers memory, password, and file-limit failures.
Decide what to archive and what to share
Page operations create a new derivative file; they do not replace your source of record. Give the derivative a name that states its purpose, such as policy-handout-pages-1-3-8-10.pdf, and keep a short note of the source filename and ranges used. That matters when a reviewer asks why an appendix is missing or when a later revision changes the page numbers. For records or legal documents, check whether bookmarks, signatures, accessibility tags, and embedded attachments are required before choosing a browser tool. A visually correct five-page handout can still be unsuitable if the workflow depends on those document features.