Delete PDF Pages
Remove pages from a PDF using a visual thumbnail grid. Bulk select, invert, keep only what you need. 100% in your browser.
Delete PDF Pages
How to delete PDF pages
Add your PDF
Drag your file onto the dropzone or click to choose. The tool renders a thumbnail for every page so you can see what you're keeping and removing.
Click pages to mark them
Click any thumbnail to mark it for deletion — the card turns red and the page number gets struck through. Click again to unmark. The counter at the top tracks how many pages will be kept.
Use bulk actions when needed
Select all marks every page. Clear removes every mark. Invert swaps the kept and deleted sets — useful when it's easier to mark the pages you want to keep, then invert.
Click Delete and download
pdf-lib builds a new PDF containing only the kept pages and saves it as {original}-trimmed.pdf. The kept pages appear in their original order — this tool doesn't reorder. (For reordering, use our PDF Reorder Pages tool.)
Deleting PDF pages — a rebuild, not an edit
Removing a page from a PDF is not, technically, "deleting" anything. The PDF file format (ISO 32000) is structured around an object graph and a cross-reference table; pages hang off a page tree and reference shared resources — fonts, images, form definitions — that may also be used by other pages. Truly deleting a page means building a new document that contains only the kept pages: each kept page object is copied, with everything it references, into a fresh file, and a resource shared by two kept pages is copied once. That is what pdf-lib's copyPages() does, and what this tool runs locally in your browser tab. Anything that lives at document level rather than page level — the bookmark outline, the AcroForm dictionary that makes form fields fillable, the Title and Author — is not part of any page and therefore does not come across.
Why "redact" and "delete" are different
Online PDF editors often confuse "delete" with "redact." A redaction overlays a black rectangle on top of content but leaves the underlying text intact — forensic tools can recover the original. A deletion removes the page from the document entirely, so there is no underlying content to recover. For sensitive workflows — legal discovery, HR records, financial filings — this distinction is critical. If you actually want the content gone, you delete the page; if you need the page but not a passage, use a true redaction tool (our PDF Redactor) and then check the result with the PDF Redaction Checker. This page tool deletes. There is no recovery.
Deleted pages are gone from the output — not hidden, not crossed out. That is because this tool writes a new file and copies only the kept pages into it; the bytes of a deleted page are never written.
What this tool does — and what it doesn't
This tool removes selected pages and outputs a new PDF containing the remaining pages in their original order. It preserves text content, embedded fonts, images and the link annotations on kept pages (links pointing to deleted pages become dead, which is unavoidable). Form fields on kept pages keep their appearance — you will see the filled-in values — but the new document has no AcroForm dictionary, so they stop being fillable fields. Bookmarks are not copied. Document metadata is not copied either: the output's Producer reads "RECATOOLS Delete PDF Pages" and its creation date is the moment you clicked. What it does not do: reorder pages (use PDF Reorder Pages), redact content within pages (PDF Redactor), handle encrypted files (see the FAQ), or recover deleted content — keep a backup of the original.
What a rebuilt PDF keeps, what it drops, and why the file barely shrinks
"Deleting" a page in a PDF means rebuilding the document with only the kept pages. The original file is unchanged — the deletion lives in the output.
Forensic tools cannot recover deleted pages from the output PDF — the bytes that contained those pages are not in the file at all.
Bookmarks (the outline shown as a sidebar in Acrobat) are a document-level tree, not part of any page, so copyPages() does not bring them along: the output has no bookmarks at all.
Form fields on kept pages keep their drawn appearance, values included, but the AcroForm dictionary that defines them is document-level and is not copied — in the output they are pictures of fields, no longer fillable.
Hyperlinks WITHIN the kept pages are preserved. Hyperlinks pointing to a deleted page become dead — that's a fundamental limitation, not a bug.
The output file size is usually smaller than the input but not in direct proportion. Shared fonts and images that survive on kept pages keep their full byte cost.
PDF does have a recoverable state — an incremental update appends changes and leaves the old objects in the file (ISO 32000 §7.5.6). This tool never uses it: the output is a fresh file, so nothing deleted is present to recover.
You can delete every page except one — but not every page. This tool refuses to write a document with no pages: the button disables the moment nothing is left, and "Select all" is really a shortcut for "Select all, then un-mark the ones to keep".
The pdf-lib library doing the rebuild (v1.17.1, 525 KB minified, MIT-licensed) runs entirely in the page; its own documentation says it does not support encrypted PDFs, which is why the encryption FAQ below says what it says.
The output's Producer is set to "RECATOOLS Delete PDF Pages" and its creation date to the moment you clicked; the source file's Title, Author and other Info fields are not copied, so the trimmed file is also metadata-clean.
FAQ
-
No. The pdf.js library renders thumbnails in your browser; pdf-lib applies the page deletion in your browser. Your PDF is read into memory, modified in memory, saved back to your downloads folder. Open DevTools → Network and watch — zero outbound traffic.
-
No. This tool will not write a document with no pages — the Apply button is disabled when zero pages remain. You can delete every page except one.
-
Yes — in the output file. The deleted pages are not present in the saved PDF at all. The original input file on your computer is unchanged; you can re-add the deleted pages by re-loading the original. If you didn't keep a backup, the deletion is permanent.
-
It flips your selection — pages currently marked for deletion become kept, and pages currently kept become marked for deletion. Useful when it's easier to mark the pages you want to keep, then invert: e.g., 10-page document where you only want pages 3 and 7 — click those two, then click Invert.
-
Partly. Form fields on kept pages keep their appearance and visible values, but the document-level AcroForm dictionary is not copied, so they are no longer fillable in the output. Bookmarks are not preserved — the output has no outline. Link annotations on kept pages are preserved; links pointing to deleted pages become dead.
-
Deletion removes the page entirely — no content remains in the output. Redaction overlays the visible content (typically with black rectangles) but the underlying data may still be recoverable by forensic tools. For genuinely sensitive content, deletion is the safer choice. Redaction is for cases where you must preserve the page (for context) but hide specific content.
-
Almost always, but not always proportionally. Shared resources (fonts, images) on kept pages stay in the file at full cost. Deleting 50% of pages might only shrink the file by 30% if the shared resources are the dominant byte cost.
-
No — and this includes "restriction-only" files (owner password, no-print/no-copy), whose content streams are still encrypted. pdf-lib 1.17.1 does not support encrypted PDFs; the tool now refuses them at load (it used to accept them without an error), because pdf-lib's own documentation warns that modifications to such a document "may fail, or have unexpected results" — pages copied out of an encrypted file typically render blank or garbled. Remove the encryption first (Acrobat, or export a copy from macOS Preview, or
qpdf --decrypt), then delete pages here. -
No hard limit other than browser memory. Thumbnails render one page at a time, so the preview of a several-hundred-page document takes noticeably longer than the rebuild itself, which copies only the kept pages. On a phone, long documents are limited by memory before anything else.
-
Yes on iOS Safari and Chrome on Android. The thumbnail grid reflows to one or two columns on phones. Tapping a thumbnail to mark for deletion works the same as desktop clicking.
Method & sources
How it computes
Renders a thumbnail of every page with pdf.js 4.10.38, lets the reader mark pages, then builds a new document with pdf-lib 1.17.1: PDFDocument.create(), copyPages(source, keptIndexes) and addPage() for each copy, saved as {name}-trimmed.pdf. Because the output is a fresh file rather than an edited copy, a deleted page's objects are never written; page-level content (text, fonts, images, link annotations) travels with each copied page, while document-level structure —
What this tool implements
- Pages are copied as page-tree leaves (ISO 32000-2 §7.7.3): the copier deep-copies each page dictionary and everything it references except its Parent, so a font or image shared by two kept pages is written once.
- Output order is the input order — the kept indexes are ascending; reordering is a different tool.
- The tool will not write a document with no pages: the Apply button disables when zero pages remain, and applying with nothing marked is a no-op with a message.
- Encrypted input is not supported. The tool passes ignoreEncryption:true so the file loads, but pdf-lib's documentation states this 'does not decrypt the document' and that modifications 'may fail, or have unexpected results' — the page says so and tells the reader to decrypt first.
- Form fields on kept pages keep their appearance streams but the output has no AcroForm dictionary, so they are no longer interactive; bookmarks are not copied; the source Info dictionary is not copied and the output Producer is set to 'RECATOOLS Delete PDF Pages'.
- Thumbnails render sequentially at 160 px width via pdf.js getViewport/render; a render that does not resolve is raced against a timeout so a stuck page cannot block the grid.
Sources
- ISO 32000-2:2020. Document management — Portable document format — Part 2: PDF 2.0 (Edition 2, confirmed 2026) — §7.7.3 page tree, §7.5.6 incremental updates, §12.5 annotations, §12.7 interactive forms (AcroForm). https://www.iso.org/standard/75839.html
- ISO 32000-1:2008 (PDF 1.7), full text as published by Adobe. https://opensource.adobe.com/dc-acrobat-sdk-docs/pdfstandards/PDF…
- Hopding A. pdf-lib v1.17.1 (6 November 2021) — PDFDocument.load options (ignoreEncryption), copyPages(), and the README section 'Encryption Handling'. https://github.com/Hopding/pdf-lib/releases/tag/v1.17.1 and https://pdf-lib.js.org/docs/api/classes/pdfdocument
- Mozilla. pdf.js v4.10.38 (31 December 2024) — getDocument, getPage, getViewport, render. https://github.com/mozilla/pdf.js/releases/tag/v4.10.38 and https://mozilla.github.io/pdf.js/api/
What can make this go out of date
- none at runtime — both libraries are vendored copies (byte-identical to the audited npm packages); a pdf-lib release that adds encryption or XMP support would change what the FAQ can promise.
Abridged — the full review record for this tool runs longer than the list above.
Pick up where you left off
Stored only in this browser — never sent to our servers.