PDF Extract Pages
Pick exactly the pages you need and download them as a brand-new PDF.
100% client-side — your files never leave your browser.
Drag & drop your PDF here
or click to browse files
How do you extract pages from a PDF? Extracting means copying a chosen subset
of pages into a brand-new PDF while leaving the original file untouched. This tool reads the
document structure with pdf-lib, copies each selected page object — text,
fonts, images and vector artwork together — into a fresh document, and saves it as
extracted-pages.pdf. Pages always come out in ascending order, whatever order you
picked them in.
How to Extract Pages from a PDF
-
Load the PDF — Drag a file onto the upload zone or click to browse. The document is read into memory with
FileReader, its page count is measured, and one numbered tile appears per page. - Pick pages by clicking — Click any tile to add it to the selection and click again to drop it. The counter above the grid shows how many pages are currently selected, and the extract button stays disabled until at least one is.
-
Or type a page range — Enter something like
1-3, 5, 7-10and press Apply. Numbers are 1-based, anything outside the document is ignored, and applying a range replaces the current selection rather than adding to it. -
Extract the new document — Click Extract Pages. Each selected page is copied into a new PDF in ascending page order and the file downloads as
extracted-pages.pdf. - Check the result — Open the download and confirm the page count. Cross-references that pointed at pages you left behind will no longer resolve, so re-check any internal links or a table of contents.
How Page Extraction Works
A PDF is not a stack of independent sheets. It is a graph of numbered objects — page dictionaries, font programs, image streams, colour spaces, annotations — tied together by indirect references and indexed by a cross-reference table at the end of the file. A page typically owns very little on its own; it points at shared resources that other pages point at too. That is why you cannot extract pages by cutting bytes out of the file with a text editor.
The extraction here uses pdf-lib. The uploaded bytes are parsed into an
in-memory object graph, an empty destination document is created, and for each page you
selected copyPages walks that page's references and deep-copies everything it
depends on into the new document, renumbering objects as it goes. The copied page is then
appended with addPage. Finally save serialises the new graph and
writes a fresh cross-reference table, producing a standalone, valid PDF.
Because the copy is object-level rather than pixel-level, nothing is re-rendered and nothing is re-compressed. Text stays selectable, images keep their original resolution and encoding, and vector artwork stays vector. The extracted file is usually much smaller than the original, since resources used only by the pages you dropped are never copied across.
Page Range Syntax
| You type | What it selects |
|---|---|
5 | Page 5 only |
1-3 | Pages 1, 2 and 3 |
1-3, 5, 7-10 | Pages 1, 2, 3, 5, 7, 8, 9 and 10 |
10-7 | Pages 7 to 10 — reversed ranges are normalised |
2, 99 | Page 2 only, if the document has fewer than 99 pages |
5- | Nothing — open-ended ranges are not supported |
Numbering is 1-based and matches the page labels your PDF reader shows in the page-number box, which is not always the number printed on the page itself. If a book has roman-numbered front matter, count from the very first sheet in the file.
What Carries Over into the Extracted File
| Element | Survives extraction? |
|---|---|
| Text, fonts and character encoding | Yes, copied unchanged and still selectable |
| Images and vector graphics | Yes, at original resolution, not re-compressed |
| Page size, rotation and orientation | Yes, taken from the source page |
| Page annotations on a copied page | Usually, since they hang off the page object |
| Bookmarks and the outline tree | No — these live at document level |
| Interactive form field registration | No — expect fields to lose interactivity |
| Internal links to pages you excluded | No — the destination no longer exists |
| Document metadata and attachments | No — the new document starts empty |
Encrypted PDFs are refused. The parser rejects any document carrying encryption, including files that only have an owner password restricting editing. Remove the protection in the application that created the file, or open it in a reader and re-save an unprotected copy, then extract from that.
Frequently Asked Questions
It means building a new PDF that contains only the pages you chose, while the source file stays exactly as it was. Extraction is a copy operation, not a cut: nothing is deleted from the original, and the result is a complete, standalone document that any reader can open without the file it came from.
The page content is copied at the object level, so fonts, text, images, vector artwork, page size and rotation all come across unchanged and nothing is re-rendered or re-compressed. What does not survive is document-level structure: bookmarks, the interactive form catalogue, metadata, and internal links whose target page you left out.
Separate entries with commas and use a hyphen for a span, mixing the two freely — 1, 3-5, 8 selects pages 1, 3, 4, 5 and 8. Numbering starts at 1, a reversed span such as 9-6 is normalised, and page numbers beyond the end of the document are silently skipped. Applying a range clears whatever was selected before it.
No. The parser refuses any encrypted document, and that includes files carrying only an owner password that restricts editing rather than opening. Unlock the file first — open it with the password in a PDF reader and save an unprotected copy — then extract from that copy.
The file is read by your browser with FileReader and parsed in the page you are looking at; the bytes never travel over the network. Only the pdf-lib script itself is fetched from a public CDN. Closing the tab discards everything, so a confidential contract can be trimmed without it leaving the machine.
Extract when you want one new file holding a specific subset, such as pulling three pages of a report out for a colleague. PDF Split when you want the whole document broken into several files at once. PDF Delete Pages when the document is nearly right and you only need a handful of pages gone.
The whole file is held in memory, so the practical limit is the RAM available to the tab rather than a fixed size. Documents of a few hundred pages are routine; very large scanned files with high-resolution images can make the browser slow or run out of memory. If that happens, split the document first and extract from the smaller part.
The grid draws a lightweight numbered tile per page rather than rendering each one, which keeps a long document responsive and avoids the cost of rasterising pages you are not going to keep. If you need to see the pages before choosing, open the PDF alongside in your usual reader and use its page numbers, which match the tiles exactly.
Use Cases
Sending One Contract Clause
A contracts administrator pulls the three pages covering a payment schedule out of a 90-page master agreement so the supplier receives only what concerns them.
Course Readings
A lecturer extracts a single chapter from a scanned book PDF for a seminar pack, keeping the file small enough to attach to the course page.
Board Pack Excerpts
An analyst lifts the summary and the two appendix pages from a quarterly report so directors read the figures without wading through the full deck.
Expense Claim Receipts
An employee whose bank exports a whole month as one PDF extracts the four pages holding the relevant transactions to attach to a claim form.
Print Shop Proofs
A designer extracts a single spread from a long catalogue to send for a colour proof, avoiding the cost and upload time of the complete file.