How to Extract Pages from a PDF and Keep Only What You Need

You have a 200-page PDF. You need page 14. Forwarding the whole file to get one page out feels absurd — and if the other 199 pages contain anything sensitive, it's also a disclosure problem.

Extraction is a retrieval task, not reorganization: pick a subset of pages, write a new PDF of just those, leave the original untouched. A PDF page isn't a flattened photograph of a document — it's a discrete object referenced from the file's own page tree, which is exactly why a good extractor can lift page 14 out cleanly instead of reprinting the whole document and hoping the unwanted parts fall away cleanly.

The Fast Way

1. Open the Extract Pages tool
2. Drop your file in
3. Type a range like 1-5, 8, 12-15 or click the page thumbnails
4. Download the new file with only the chosen pages

The original never changes. A wrong pick costs nothing; run it again.

Two details worth knowing before you start. If the file is password-protected, an inline password field replaces the upload card — no modal, no separate unlock page, no re-upload. One caveat: unlocking a genuinely encrypted PDF re-renders its pages as images before copying, so the extracted pages look identical but lose their selectable text (the password section below explains why). And if you type a range the document doesn't have — page 47 in a 30-page file, or the numbers reversed as 12-8 — the tool tells you exactly which part of the range is invalid instead of silently dropping it or extracting the wrong pages.

Why the Free Workarounds Have a Caveat

Browser Print → Save as PDF with a custom page range is the usual free fallback, and on Windows the built-in Microsoft Print to PDF driver does the same trick from any application's print dialog. Both work — but print is a virtual printer, not a page copier: interactive form fields get flattened into static text, hyperlinks die, and the print driver re-encodes every page through its own rasterizer rather than copying the original page objects. Very large documents can also choke the print dialog or the driver itself.

A page-copying extractor keeps interactivity intact because it copies the actual page objects instead of reprinting them. For a tax worksheet or any fillable form, that difference decides whether the extracted pages remain usable rather than merely legible.

macOS Preview can also extract pages, by dragging thumbnails out of the sidebar into a new window, and because it operates on the page objects directly rather than through a print driver, it preserves them the same way a dedicated extractor does. It's Mac-only and easy to miss if nobody ever showed you the gesture, but it's a genuinely clean method if you're already on a Mac and don't want to open another tab.

Extract vs. Split vs. Delete vs. Organize: Pick the Right Operation

Extract and Delete can produce the same output — but when you want 2 pages out of 500, "select those two" is a sane path and "select 498 to remove" is not. Use Split PDF when every page needs to land somewhere — breaking a scanned folder of receipts into one file per receipt, say (splitting PDF pages covers that job in detail) — Delete Pages when you're pruning a handful of unwanted pages from an otherwise-intact document, or Organize PDF when every page you have is a page you want, just in the wrong order.

What Survives Extraction: Forms, Links, and Annotations

Because extraction copies page objects rather than re-rendering pixels, most of what's on the page travels with it. A hyperlink is stored as its own annotation object attached to the page it sits on, so copying the page carries the link along and it stays clickable in the new file.

Fillable form fields are the one case worth understanding before you rely on it. A field's on-page appearance — the box, and any text already typed into it — lives in the page's own annotation list and copies over with the page, so what you see stays visible. What doesn't automatically come along is the document-level form registry most PDF viewers consult to decide which boxes on the page are still live, editable controls, because that registry belongs to the whole document, not to any one page. In practice, that means text in a form field usually survives extraction as readable content, but the field itself can stop behaving as an editable input in the new file — it looks right until someone tries to click into it.

If the whole point of pulling pages out is a form the recipient still needs to fill in, test the output before you send it. And if it comes up short, consider the opposite operation instead: Delete Pages on the original keeps every remaining field fully interactive, because deleting edits your existing document in place rather than assembling a new one from scratch — nothing about the document's form registry ever gets rebuilt.

Extracting Pages from a Password-Protected PDF

Extract Pages will unlock a protected file for you inline, but it's worth knowing what happens next. The PDF library doing the actual page-copying can't decrypt a file that genuinely requires a password to open, so once you unlock one, the tool works around that by re-rendering each page through the browser's own PDF engine — at roughly 144 DPI — and rebuilding a fresh, unencrypted PDF from those images before your page selection is ever applied.

The output looks the same: same layout, same visible text, same images. What's different is that the text is no longer live text — it's pixels that happen to look like text. If the extracted pages need to stay searchable and copy-pasteable, and the source PDF genuinely needs its password to open, that's a real limitation rather than a bug in this one tool: any tool built on the same copy-the-pages approach runs into the same wall, because the encryption has to be resolved somehow before a page object can be copied at all. A PDF that isn't actually encrypted skips this step entirely — its pages copy through as the original vector and text content, completely unchanged.

The Privacy Angle

Extraction is disproportionately a sensitive-document operation — the point of "share only these pages" is usually that the other pages are confidential. Extracting one invoice from a bank statement means uploading the entire statement to a server if you use an upload-based splitter, even though whoever asked for it only needed one line item.

That mismatch is exactly what data-protection law is written to prevent. The EU's GDPR states its data minimisation principle in plain terms: personal data handling should be "adequate, relevant and limited to what is necessary in relation to the purposes for which they are processed" (Article 5(1)(c)). An upload-based extractor structurally can't honor that — it has to receive the whole document to hand you back part of it, whether or not the rest of the statement, the rest of the deposition, or the rest of the medical chart was ever relevant to why you needed page 14.

Client-side processing sidesteps that question rather than answering it with a privacy policy: no upload, no server copy, nothing transmitted that could later be breached. Open your browser's Network panel while you extract and watch — no file bytes leave the tab.

Practical Tips

Printed page numbers ≠ PDF page numbers. Front matter offsets everything: a report with a four-page cover, table of contents, and executive summary pushes the page labeled "1" in the body text to the fifth physical page in the file. The mismatch exists because a document can carry its own page-labeling scheme — Roman numerals for the front matter, then Arabic numerals restarting at 1 for the body — separate from the file's raw page order. The W3C's accessibility guidance for PDF page numbering describes this exact front-matter offset as the standard reason the two numbers diverge. Trust the viewer's physical page counter, not the number printed on the sheet. This accounts for almost every off-by-N mistake.

Mixed range syntax saves passes. 1-5, 8, 12-15 in one go beats three separate extractions, and you can freely mix single pages and ranges in the same input.

Extracting from more than one source file. If the goal is pages pulled from different documents — the cover letter from one PDF, the signature page from another — extract each range, then merge the PDF files into a single document.

Check the extract before deleting the source. Flip through the output, especially when a range was typed blind rather than picked from thumbnails — a range that's off by one page is easy to type and easy to miss until someone downstream opens the file.

If the extracted pages need to stay a fillable form, verify it before sending. As covered above, a field usually keeps showing its last value but may not stay interactive. A ten-second check now is cheaper than a confused email later.

On phones, extraction is one of the most common on-the-go tasks — sign one page of a lease in a waiting room, email it back before the appointment ends. The page grid is fully responsive and works with a tap; there's no separate mobile app to install.

Frequently Asked Questions

Does extracting pages change the original file?

No. Extraction always produces a new PDF, downloaded separately — the source file on your device is never modified or re-saved. If the result comes out wrong, discard it and run the range again against the untouched original; there's nothing to undo.

Extract vs. Split — which should I use?

Extract when you know exactly which pages you want and everything else is unwanted. Split when every page needs to end up somewhere, such as breaking a scanned batch of forms into one file per form. As a rule of thumb: if you're removing more pages than you're keeping, selecting the few you want (Extract) is almost always less error-prone than selecting everything you don't (Delete).

Do hyperlinks and form fields survive extraction?

Hyperlinks, yes — a link is its own object attached to the page, and copying the page carries it over intact. Form fields are more nuanced: the value already typed into a field stays visible because it lives in the page's own content, but the document-level registry that makes a field editable doesn't automatically travel into the new file, so a field can look right and stop functioning as an input. See "What Survives Extraction" above for why, and use Delete Pages instead of Extract if the form absolutely needs to stay fillable.

What happens if I select every page?

The tool stops you before producing a pointless file: selecting 100% of a document's pages and running Extract prompts you to use Download instead, since a full-document "extract" would just be an identical copy of the file you already have.

Can I extract pages from a password-protected PDF?

Yes — drop the file in and an inline password field appears before the page grid does. The trade-off is described above: because the underlying library can't decrypt the file directly, unlocking it re-renders every page as an image before copying starts, so the extracted pages look the same but lose their live, selectable text. A PDF that isn't password-protected never goes through this step.

Does extraction work on scanned PDFs?

Yes, and it comes through cleanly either way. A scanned page's content is an embedded image rather than live text, and copying a page carries over whatever is actually on it — image, text, or both — without needing to know which. If the scan has already been run through OCR and carries an invisible text layer under the image, that layer copies over too.

Is there a limit to file size or page count?

There's no fixed cap built into the tool, but everything runs in your browser rather than on a server, so an unusually large file — many hundreds of pages, or a heavily scanned document — leans on your device's memory instead of someone else's, and will take longer than a small one.

Can I extract pages on my phone?

Yes. The page grid is responsive down to a two-column layout on small screens, selection works with a tap, and there's no app to install — the same browser tab that opened the file produces the download.

Extract the pages you need here. The file stays in your browser.