PDF Page Remover

PDF Page Remover does not hide or flag pages - it rebuilds the document from scratch with pdf-lib. The removePages() routine loads your file, computes the complement of your selection (every page index that is NOT marked for removal), then creates a brand-new PDFDocument and copyPages() each surviving page into it in order before calling save(). The result you download (named {original}_removed_pages.pdf) is a standalone, valid PDF containing only the kept pages, not the original with content suppressed. Because the page tree is reassembled, the structural page index shifts: a page that was #8 may become #6. This has real consequences - pdf-lib's copyPages does not carry over the document outline, so internal bookmarks or hyperlinks that targeted a removed page become dead, and any page numbers printed INTO the content itself (a footer reading 'Page 7') are baked pixels/text and are NOT renumbered.

Selection is flexible and happens entirely client-side. You can click individual page tiles to toggle them, or type a range string like '1,3,5-7' into the Page Range box - the parser splits on commas, expands hyphenated ranges, and silently clamps anything above the document's page count or at/below zero (the loop guards with i <= pageCount and i > 0). Quick-select buttons cover the common cases: All, None, Invert (flip the current selection), Odd, and Even, which makes dropping every blank back-of-sheet scan a single click. A live counter shows Total Pages, To Remove (red), and Remaining (green), so you always see the resulting page count before committing. Two guards run when you press Remove: you must select at least one page (the button is disabled at zero selection), and you cannot remove all of them - if your selection equals or exceeds the page count, removePages aborts with a 'Cannot remove all pages' toast and a yellow 'At least one page must remain' notice is shown inline.

Page previews are rasterized in the browser with pdfjs-dist (the worker at /workers/pdf.worker.min.mjs), rendered at scale 0.25 and encoded as JPEG at quality 0.5 to keep the grid lightweight. Crucially, thumbnail generation is capped at the first 30 pages via Math.min(count, 30); pages beyond that still appear as selectable tiles but show a numbered placeholder instead of an image, so on a very large document you select the later pages by number rather than by sight. You can switch between a grid view (a dense thumbnail mosaic, 5 to 8 columns) and a list view (one row per page with a small preview and a 'Will be removed' label). As a decision rule: reach for Page Remover when you want to drop a few scattered pages from a document you otherwise keep; reach for the extraction tool when you instead want to pull pages OUT into a new file.