Remove PDF pages

Remove PDF Pages creates a new PDF from every source page except the ones you select. The original file is never edited in place. If a 20-page document has blank pages 4 and 11, entering `4,11` produces an 18-page copy whose remaining pages stay in their original order. The output positions change after each deletion, but the selector always refers to positions in the uploaded source, not to the renumbered result.

The tool copies whole retained pages. It does not erase an object from part of a page, crop margins, hide an annotation, or remove one attachment. That distinction matters for privacy. If page 7 contains both permitted and restricted information, deleting some other page does nothing to the content on page 7. Partial-page removal requires proper redaction or editing software.

All selected pages are removed once, even if a number appears more than once. The result must retain at least one page. A request that removes every source page is rejected because an empty PDF would not be a useful output.

Use one-based page positions separated by commas. A single number removes one page, and a hyphen describes an inclusive range. `2,5-7,14` removes pages 2, 5, 6, 7, and 14. Spaces are accepted. Reversed ranges are normalized, so `9-6` removes pages 6 through 9. Invalid tokens, zero, negative values, and numbers beyond the document's page count trigger an error.

The position shown by the PDF viewer may differ from the number printed on the paper. Front matter often uses roman numerals, and scanned packets may include an unnumbered cover. If the footer says page 10 while the viewer says 13 of 60, the selector expects 13. Confirm each target by looking at the viewer counter and distinctive page content. Do not rely on footer numbers alone.

A useful check is to write down the pages being removed and the pages immediately before and after them. For a duplicate at source position 32, note what should end page 31 and what should begin page 33. After processing, those neighbors should sit next to each other. This boundary method catches the most common mistaken deletion quickly.

Scanner batches often contain blank backs, separator sheets, or double-fed duplicates. Review thumbnails first because a nearly blank page may carry a faint signature, page number, barcode, or continuation mark. Increase contrast in a suitable viewer if needed. Only select a page after confirming it has no content the record must retain.

For an outdated appendix, identify the complete section boundaries. A heading may start on the previous page, and footnotes may continue onto the next one. Removing pages 40 through 47 from a policy manual can leave a table of contents entry, cross-reference, or bookmark that still promises the deleted appendix. Page removal does not rewrite those references. Update the source in its authoring application when the document must read as a newly published edition.

Accidental duplicates are best checked side by side. Similar headers can make two valid pages look identical in thumbnail view. Compare body text, dates, running numbers, and signatures. After deletion, download the copy and verify both sides of every join before distributing it.

Meeting packets often need a different kind of cleanup after circulation. Draft agenda pages, superseded proposals, or confidential closed-session material may need to leave a public copy, but the published minutes must still explain what was considered. Follow the records schedule and approval process rather than deciding from appearance alone. If a removed page is cited elsewhere, replace the citation with an approved note in the source application. A silent gap can mislead readers and make later records requests harder to answer.

Calculate the expected page count before processing. A 42-page source with selections `3,8-10,30` should produce 37 pages because five unique pages are removed. If ranges overlap, count unique pages rather than adding both ranges blindly. For example, `4-8,7-10` removes pages 4 through 10, which is seven pages, not nine.

Open the output and inspect thumbnail sequence. Check the first page, last page, and each deletion boundary. Search for a phrase that appeared only on an unwanted page, but remember that scans may not be searchable and hidden content can escape a simple query. Also search for required section titles to confirm they remain. If the file is a formal record, compare it with the approved removal list and have another reviewer sign off.

Test links, bookmarks, forms, and printed page references. A bookmark can point to a deleted destination. A sentence such as `see page 18` can now direct the reader to different content because physical positions shifted. The tool does not renumber visible footers or repair cross-references.

Each input must be a readable PDF smaller than 100 MB. Encrypted files are refused; the browser does not prompt for a password or ignore protection. A damaged cross-reference table, unsupported object, or mislabeled non-PDF file can also fail during loading. If you are authorized to edit the document, export a fresh working copy from its source application and test that copy while retaining the original.

Deleting pages rebuilds the PDF from retained page objects. Advanced document behavior may not carry over exactly. Digital signatures can become invalid, form scripts may reference removed fields, internal links may lose destinations, and document-level attachments or layers may not behave as expected. This tool does not flatten, sanitize, or certify the result.

Browser memory can become the limiting factor before the 100 MB cap, particularly for image-heavy scans on mobile devices. Close other large tabs and use a desktop browser if processing stops. If the job remains too large, use an approved local PDF application. Avoid moving confidential files to an unknown upload service simply because the browser ran out of memory.

Compatible PDFs are processed on the device and are not sent to a PWRKIT processing server. The new copy is still saved through the browser, often into a Downloads folder that may sync or be backed up. Use a managed location when the file contains client, employee, medical, financial, educational, or legal information. Clear temporary copies according to your retention policy after the accepted output is filed.

Removing a page is not the same as redacting a fact. The same name or account number may appear in headers, bookmarks, annotations, form data, metadata, attachments, or another page. A deletion checklist should include searches for the restricted information and a visual scan of nonsearchable content. For high-risk disclosure, use software that can sanitize hidden content and produce an audit record.

Preserve the source and the removal instruction when the change affects a record. A filename such as `minutes-public-copy.pdf` tells more truth than `minutes-final.pdf` if pages were excluded for publication. Record which source revision was used, which page positions were removed, who approved the removal, and how the output was checked.

Deletion can break the logical flow for people who navigate by headings, bookmarks, links, or screen-reader structure. A retained paragraph may begin with `as described above` after the referenced page is gone. A heading level may start at level three because its parent section was deleted. Review the content around every new join, not only its visual spacing.

Run an accessibility checker on the output when it will be published or used with assistive technology. Confirm the document title, language, reading order, tagged headings, link destinations, and form labels. This page tool does not add OCR to scanned pages, repair tags, or write alternative text. Fix accessibility in an authoring or PDF remediation application when the source does not meet the requirement.

Keyboard-test interactive material. Tab through retained fields and links, and make sure focus does not jump to a removed page or disappear. If visible page numbers changed but the printed labels did not, provide an accessible note that explains how to cite locations in the edited copy.

Removal is clearest when most of one PDF should stay and a short list should go. Extraction is clearer when only a small subset should become a new document. From a 100-page manual, keeping pages 12 through 18 is easier to express as an extraction than as a removal of pages 1 through 11 and 19 through 100. The result can contain the same pages, but the instruction is easier to review.

Split PDF creates one separate PDF per source page and bundles them in a ZIP. It does not delete selected pages from a consolidated copy. Merge PDF joins several files. Rotate PDF changes the orientation of every page, while Compress PDF seeks a smaller structural representation. None of these jobs removes part of a page or performs secure redaction.

When several edits are required, make content decisions before size optimization. Remove approved pages, verify the sequence and context, then rotate or compress the accepted copy as needed. Keep filenames that expose each stage. The untouched source remains the point of reference if a reviewer later questions a missing section.

Do not use deletion to conceal a revision from collaborators who are entitled to the full history. A review copy, publication copy, and archival master can legitimately contain different pages, but each needs an honest label and access rule. If a page was removed because it was erroneous, correct the source record through the normal amendment process. The browser output cannot document why a page disappeared or who authorized the change.