PDF Page Deleter

Source PDF:

Choose one unlocked PDF up to 75 MiB and 700 pages, or create the safe eight-page sample.

Pages to delete:

Examples: 2, 5-7, 10-, odd, or even. At least one page must remain.

Output filename:

Optional. Unsafe filename characters are replaced before download.

Cleaned PDF ready

{{ computation.values.kept_count }} pages kept

Source pages: {{ computation.normalizedInputs.page_count }}
Deleted pages: {{ computation.values.selection_label }}
Kept pages: {{ computation.values.kept_count }}
Output size: {{ formatBytes(outputSize) }}

Rewriting can invalidate digital signatures and affect bookmarks, named destinations, forms, page labels, or tagged structure. Open and review the downloaded copy before sharing it.

Source page	Action	Output page	Review note	Copy
{{ row.sourcePage }}	{{ row.action }}	{{ row.outputPage }}	{{ row.note }}

Removing pages from a PDF creates a new document whose remaining pages keep their original order. The important distinction is between a page's position in the source file and its new position after deletion. If source pages 2 and 5 are removed, later pages shift forward even though their visible content stays the same.

Page position is not always the same as the number printed on the page. A report may begin with an unnumbered cover, use Roman numerals for front matter, or carry custom PDF page labels. A request to remove “page 3” therefore needs a clear reference: usually the third page shown by the PDF viewer, not the printed number in a footer.

PDF page concepts that affect deletion
Document feature	What deletion can change
Page positions	Every kept page after a deletion receives a lower output position.
Bookmarks and named destinations	Links may point to removed content or lose the intended destination after rewriting.
Forms and tags	Interactive fields, reading order, or tagged structure may not survive a rewrite exactly as expected.
Digital signatures	Changing the document bytes can invalidate an existing signature.

Deletion is appropriate when the unwanted material is a complete page, such as a blank scan, duplicate insert, outdated appendix, or separator sheet. It is not redaction: removing one page does not search other pages, attachments, annotations, forms, or metadata for the same sensitive information. Always inspect the new file rather than assuming that a correct page count proves a correct document. Check the first and last kept pages, a page immediately before and after each removed range, navigation links, forms, and signatures that matter to the handoff.

How to Use This Tool:

Identify unwanted pages by their source positions before creating the cleaned copy.

Choose one unlocked PDF under 75 MiB with no more than 700 pages. The page count appears after the browser reads the file.
Enter source positions in Pages to delete. Use a single page such as 2, an ascending range such as 5-7, an open range such as 10-, or the keywords odd and even.
Review the ready message and any range warning. A range extending past the document is clipped to the final page, and duplicate selections are counted once. Descending or out-of-bounds starts must be corrected.
Select Delete selected pages. The cleaned PDF is created in browser memory only after the selection leaves at least one page.
Compare Page plan with the source, then open the downloaded PDF and inspect content, navigation, forms, tags, and signatures before sharing it.

Interpreting Results:

The kept and deleted counts confirm the selection size. The Page plan is the stronger check because it maps every source position to either Delete or its new output position.

A dash in the output-page column means that source page is absent from the cleaned copy.
Kept pages remain in source order and are renumbered consecutively from 1.
A smaller or larger byte size is not a quality score. Rewriting PDF objects and streams can change file size even when the intended pages are retained.
A successful download does not certify bookmark, form, signature, page-label, or accessibility fidelity. Open the result in the receiving viewer and test the features the document depends on.

Technical Details:

Page selection uses one-based source positions. The deletion expression becomes a unique, ascending set before the document is changed, so overlapping ranges and repeated page numbers never delete the same position twice.

Rule Core:

PDF page deletion expression rules
Expression	Selection rule	Boundary behavior
`n`	Delete source page n	n must be from 1 through the source page count
`a-b`	Delete every page from a through b, inclusive	The range must count upward; an end beyond the file is clipped with a warning
`a-`	Delete page a through the final source page	a must exist
`odd` or `even`	Delete positions with the named parity	Parity applies to source positions, not printed page labels
`all` or `*`	Select every source page	Rejected because at least one output page must remain
Comma list	Union all valid tokens	Duplicates are counted once and reported as a warning

Formula Core:

Let N be the source page count and D the unique set of deleted source positions. The number of kept pages is the source count minus the size of that set.

K = N - | D |

For each kept source page s, its output position equals s minus the number of deleted positions before it. A deleted source page has no output position.

o (s) = s - | {d \in D : d < s} |

With an eight-page source and 2, 5-7, the deleted set is {2, 5, 6, 7}. The kept source pages are 1, 3, 4, and 8, which become output pages 1, 2, 3, and 4 respectively.

Transformation Core:

The source PDF is loaded without ignoring encryption. Selected pages are removed from highest position to lowest so earlier indexes do not shift during deletion. The remaining document is then serialized as a new PDF with object streams enabled.

This is a structural rewrite, not a visual rasterization of every page. It preserves the retained page objects where the PDF writer can, but the surrounding document catalog, references, forms, labels, bookmarks, tags, and signatures still require post-export inspection.

Privacy and Document Integrity:

The source and cleaned PDF remain in the current browser tab; this page does not upload the document. Closing or refreshing the tab discards its in-memory copy unless you downloaded the result.

Local processing does not make deletion a secure redaction method. Sensitive text may remain elsewhere in the document, and the new file may retain or alter interactive structures in ways that are not visible in a page-count check. Use a dedicated redaction and sanitization workflow when the goal is irreversible removal of confidential content.

References:

ISO 32000-2:2020, Document management — Portable document format — Part 2: PDF 2.0, International Organization for Standardization, 2020.