FineDocs
English

Delete pages from a PDF

Remove unwanted pages from a PDF free, with no page limit or watermark. Blank scans, cover sheets and duplicates go without re-uploading the file.

Drop a file here

or paste from the clipboard

Options

These pages are removed and everything else is kept. At least one page must remain.

Results appear here. Nothing is uploaded.

    How it works

    Your selection is resolved into page indices, then removed with @cantoo/pdf-lib in descending order. That ordering detail matters: deleting page 2 first would renumber every page after it, so ascending removal quietly deletes the wrong pages. Working from the end means each removal cannot disturb the indices still to be processed. Remaining pages keep their original content streams untouched, so nothing is re-encoded and no wasm is needed.

    What to watch out for

    Deleting pages frees less space than expected

    Fonts and images are shared resources referenced by many pages. Removing a page drops its references but the resource stays if anything else still uses it, so cutting half the pages rarely halves the file size.

    The table of contents does not renumber itself

    Outline entries and internal links point at specific pages. Deleting pages leaves those references pointing at the wrong place or at nothing, and no automatic repair is possible without rewriting the document's structure.

    Page numbers printed on the page stay as they were

    Numbers drawn into the page content are just text. Delete page 4 and the sheet that follows still reads "5". Only the actual document order changes.

    Limits

    Files up to about 200 MB work reliably. Beyond that, the ceiling is your device's memory rather than any rule we impose — the work happens on your own machine. Tested in Chrome 140, Firefox 143, Safari 18.

    Delete or extract — the same job from opposite ends

    Deleting pages and extracting pages produce the same result by different routes, and picking the right one saves effort.

    Deleting is the better choice when you want most of the document. A forty-page scan with three blank sheets is easier to describe as “remove 12, 19, 33” than as a list of everything to keep.

    Extraction wins when you want a small part of something long. Pulling one section out of a hundred-page report is a short selection; the equivalent deletion list would be enormous.

    The indexing detail that catches implementations out

    This is worth explaining because it is a real bug in some tools and completely invisible when it happens.

    Ask to delete pages 2 and 5. If a program removes page 2 first, the document immediately renumbers: the old page 3 becomes page 2, the old page 5 becomes page 4. Deleting “page 5” next removes what used to be page 6. The output has the right page count, contains the wrong pages, and nothing announces the mistake.

    Removing from the end backwards avoids this entirely, because deleting a later page cannot change the position of an earlier one. That is what this tool does, and a test asserts it specifically on a selection where the two orderings differ.

    Why the file does not shrink much

    A PDF is not a stack of independent pages. An embedded font is stored once and referenced wherever it is needed; the same is true of a repeated logo.

    Deleting a page removes its own drawing instructions and its references, but a shared resource only becomes unnecessary once nothing at all points to it. Cutting twenty pages out of a forty-page document therefore typically saves far less than half the size, because the fonts every remaining page still uses have to stay.

    Genuine size reduction means resampling images and subsetting fonts, which is a different operation. Every high-quality engine that does it well is licensed AGPL, so rather than ship a copyleft dependency or a misleading half-measure, that tool is not offered here yet.

    One page must survive

    Selecting every page is rejected rather than producing a zero-page file. Such files are technically malformed and most viewers refuse to open them, so the error prevents a result that would only look like it worked.

    Your files never leave your browser. Last updated 2026-07-31.