Skip to content
ToolsOnDuty - free online tools
PDF ToolsNovember 6, 2026

How to Split a PDF and Extract the Pages You Need

Select PDF pages visually or by range, extract one document and compare that workflow with exporting every source page in a ZIP.

By ToolsOnDuty Editorial Team

Split PDF workspace selecting source pages two, three and five from a five-page demonstration.

Splitting a PDF can mean two different things: extracting several selected pages into one new document, or creating a separate file for every original page. Choose the output you actually need before clicking a download button, particularly when the source contains pages that should not be distributed together.

This guide uses the ToolsOnDuty Split PDF workspace with an original five-page demonstration document. We verified the selected output's page count and source-page dimensions, inspected the ZIP's five individual PDFs and tested how page-range normalization behaves. No private client documents were used.

Open the source and confirm its page count

Choose a PDF you own or are authorized to process. Wait for the page count and thumbnail gallery to appear. Our fixture had five labeled demonstration pages, and the source information reported 5 pages ready. The document is processed in the browser rather than uploaded to a splitting service.

A source-page number is its position in the PDF, starting at one. It may differ from a page number printed inside the document, especially when a cover, contents section or appendix precedes the main numbered text. Use the gallery and document content to identify the intended pages before entering a range.

Select pages with a range or the visual picker

Enter 2-3,5 in Pages to extract. The range includes source pages two and three, and the comma adds page five. The gallery marks those three thumbnails as selected. You can also click thumbnails to toggle individual pages; the range field and selection count reflect the selected set.

Use Select all to mark the whole document or Clear to remove the current selection. Avoid assuming that a file is ready for extraction simply because thumbnails are visible. The current selection must be nonempty, and the displayed count should match what you intend to include.

Range 2-3,5 selects three PDF thumbnails while source pages one and four remain unselected.
Actual range selection: source pages 2, 3 and 5 are marked for one extracted PDF.
text
Source document: 5 pages
Pages to extract: 2-3,5
Selected source pages: 2, 3, 5
Expected extracted document: 3 pages

Understand normalization before relying on a complex range

The parser builds a set of valid source pages and sorts them in ascending order. Duplicates do not create duplicate output pages, and a reversed range is normalized rather than used to reverse the document. This tool is not a custom page-reordering interface.

We tested 5,3,2,2,99 against the five-page source: the selection remained pages two, three and five. The duplicate was removed and page 99 was ignored. Invalid fragments do not necessarily reject an otherwise valid selection, so independently check the gallery and count instead of assuming the entire input was accepted exactly as written.

Extract one new PDF containing the selected pages

Click Extract pages to a new PDF for a single output containing the selected set. Our actual export was 1,816 bytes and contained three pages. Its page dimensions matched source pages two, three and five in that order, confirming that the intended positions were copied.

Open the downloaded file and review each retained page, including its first and last page. The source is not overwritten by the extraction. Check the filename, recipient and content before sharing, and do not interpret a small byte count as proof that the document contains the correct pages.

Use the all-page ZIP button for a different output

Split every page (ZIP) creates a separate PDF for every source page, regardless of the selected extraction range. We invoked it while only pages two, three and five were selected and verified five ZIP entries. Each entry decoded as a one-page PDF matching its corresponding source page.

The actual archive was 5,864 bytes and used names ending in page-1.pdf through page-5.pdf. We then tested Select all, shown in the second screenshot. Marking all pages is useful for extraction, but it is not required to make the ZIP button include the whole source.

Select all marks five PDF pages and the separate extraction and all-page ZIP actions remain visible.
Select all marks the complete source. The ZIP action exports every source page even when the extraction selection is smaller.

Distinguish page removal from secure redaction

Extracting selected pages does not redact text inside a retained page. If a page includes an address, account number or other sensitive detail, selecting that page still distributes its content. Cropping a preview or covering a visible area is also a different problem from verified document redaction.

Review annotations, visible content and any relevant document properties before sharing a real file. This workflow does not certify removal of every hidden resource, metadata field or interactive element. For a confidential document, use a suitable review and redaction process instead of treating successful page extraction as a comprehensive privacy guarantee.

Troubleshoot selection and protected-document issues

When we cleared the selection and tried to extract, the workspace displayed an explicit instruction to enter valid pages. If extraction fails, first confirm the selected count, the source page count and whether the PDF can be opened normally. Very large or complex documents can require more browser memory and processing time.

Protected files require the appropriate authorization and password where the existing unlock workflow applies; this is not a password-recovery guide. Keep the original and inspect downloaded outputs before deleting working copies. If you only need a few pages, choose the extraction action carefully: an all-page ZIP can include material you intended to leave out.

Apply the steps from this guide directly in your browser.

Frequently asked questions

Does Split every page (ZIP) use my extraction selection?

No. The test exported all five source pages while the extraction selection contained only pages two, three and five.

Can I reorder pages by entering 5,3,2?

No. The parser sorts the selected valid page set in ascending order. Use an appropriate reordering workflow if order needs to change.

What happens to duplicate or out-of-range page numbers?

Duplicates are removed and out-of-range entries are ignored. Check the gallery and count because partially valid input can still produce a selection.

Does extracting pages securely redact information inside them?

No. Retained pages still contain their content. Review sensitive information and use a suitable verified redaction process when necessary.