Click the pages you want to keep on a visual grid — or type ranges like 1-5, 8, 12-20. The output is one single new PDF with only your chosen pages, in original order. Nothing leaves your device.
Filoraio runs the merge directly inside your browser using a small WebAssembly engine. Nothing is uploaded, nothing is queued, and you can verify it yourself — open your browser’s DevTools, switch to the Network tab, and watch it stay quiet.
They’re read into your browser’s memory through a standard file picker.
The merge runs locally — no request leaves your device while it processes.
The combined PDF lands in your downloads folder, the same way any other download would.
No upload bar, no progress spinner waiting on a server. Works offline once the page is loaded.
From drop to download takes under a minute — even on documents with hundreds of pages. The visual grid makes cherry-picking obvious; the range input is there for when you already know the page numbers.
Drag the file onto the drop zone or click to pick one. The first thumbnail appears instantly so you can confirm the right document, and page thumbnails stream in for the rest of the document while you work.
Every page is a clickable thumbnail. Click to keep — the page gets a brand border and a checkbox badge. Click again to remove it from the selection. The counter updates live so you always know how many pages your output PDF will contain.
Use Select all / None / Odd / Even / Invert to do the bulk operations in one click. Or switch to range mode and type "1-5, 8, 12-20" if you already know the page numbers. Both views stay in sync — typing a range highlights the matching thumbnails, clicking a thumbnail updates the range string.
Press Extract pages. Filoraio copies the selected pages into a brand-new PDF, in original document order, and downloads it locally. The source file is untouched and stays in place.
If you've ever needed three specific pages out of a 200-page document, this is the workflow you'll come back to. The visual picker turns minute-long Adobe Acrobat chores into a 10-second job.
Pull a single signed addendum out of a 50-page contract bundle as its own PDF — easier to file, easier to send, no need to attach the entire agreement.
Keep just the chapters you'll actually be tested on from a 600-page textbook PDF — a smaller, focused file syncs faster to your tablet and opens in milliseconds during a study session.
Save the conclusion plus the bibliography from a long paper, or extract a single figure-bearing page to share with a co-author without forwarding the whole document.
Pull three specific invoices from a 300-page bank-statement PDF and keep them as one tidy file for the client's expense report — without exposing the rest of the statement.
Extract only the signed pages from a 40-page employee onboarding packet, leaving the policy boilerplate behind for archiving in a different system.
Carve a single deliverable section out of a master proposal PDF when a client asks to see just that part — fast, no markup, no exposing earlier draft material.
Four common reasons people specifically need to keep just a handful of pages from a longer PDF — and how the visual picker, range input, and quick actions each fit.
You need to send just the signature page of an agreement, not the whole thing. Drop the PDF, click page 47 in the thumbnail grid, press Extract. The output is a one-page PDF named like the source — clean, focused, easy to attach to an email.
You need pages 8, 23, and 41 from a 200-page bank statement for an expense reimbursement. Click each one in the visual grid (the counter shows 3 of 200 selected), press Extract, and you have a three-page PDF — none of the surrounding transactions visible.
A long paper has the abstract on page 1 and references on pages 28-32. Switch to range mode, type "1, 28-32", and a six-page PDF lands in your downloads folder with the executive summary and citations only — no full-text body to scroll past.
Someone scanned both sides of a single-sided document, leaving every even page blank. Press the Odd pages quick action, then Extract — the output PDF contains only the content pages, no blanks. (For mostly-blank-on-even scans, use Even pages and Invert if the orientation is reversed.)
Four small habits that keep your extracted PDFs predictable, especially when you're working with long source documents.
The clearest rule for choosing between the two: how many files do you need at the end? One focused PDF with cherry-picked pages → Extract. Multiple separate PDFs (one per chapter, one per range) → Split. Both tools accept the same range syntax, but only Extract produces a single output.
Click pages on the thumbnail grid to build a rough selection visually, then switch to range mode to fine-tune ("oh, that's pages 4-9 and 12 — let me just type 4-9, 12 to be sure"). Or vice versa. Both views are the same underlying selection, so neither one is the canonical version.
If you click page 5, then page 2, then page 8, the output PDF is still 2, 5, 8 — original document order. This matches what almost everyone expects. For a different page order (e.g. 5, 2, 8), use the Merge PDF tool after extracting, where you can manually reorder pages.
Scanning a single-sided document on a duplex scanner gives you blank even pages. Press Odd pages and Extract to keep just the content pages. The shortcuts work with the visual picker — you'll see the highlighted thumbnails update immediately so you can sanity-check the selection before exporting.
Side by side with the average online PDF page extractor — including the ones with millions of monthly users.
| Feature | Filoraio | Typical online PDF tools |
|---|---|---|
| Where files are processed | On your device | Uploaded to servers |
| Visual thumbnail picker | Yes — click-to-keep grid | Usually range input only |
| Quick selection shortcuts | All / none / odd / even / invert | Rare on free tier |
| Watermark on output | None | Often added on free tier |
| Account required | No | Often required for >5 MB |
| File size / page-count limit | None — your device's RAM | Usually 25 MB / 100 pages |
Quick answers to the things people ask most often before using this tool.
Both tools select pages by range, but the outputs are structurally different. Extract gives you one single new PDF containing only the pages you picked, in original order — perfect for "I need just pages 5, 12, and 47-50 from this document". Split gives you several separate PDFs, one per range you defined — perfect for "break this 200-page document into chapters". If you want one focused file at the end, use Extract; if you want many smaller files, use Split.
Yes, completely free — no signup, no email, no daily quota, no premium tier, and no watermarks on the output. The only practical limit is your device's available memory; the tool itself doesn't impose a page-count or file-size cap.
No. The entire operation runs in your browser using a small WebAssembly module. Your PDF is read, its pages are copied locally, and the new PDF is saved through your browser's standard download dialog. You can verify this in DevTools → Network before pressing Extract — there's no file POST to anywhere.
When you can see the pages, you select faster and you make fewer mistakes — especially on long documents where you'd otherwise have to scroll through the PDF in another window to figure out which page numbers you want. For one-off jobs the visual grid is much quicker; for repeated jobs ("always pages 1, 5, 10") the range input is more efficient. Both views work on the same selection, so you can mix freely.
Yes — click a single thumbnail (or type its number in the range input) and press Extract. The output is a one-page PDF with that page's exact content, dimensions, and embedded objects. Very common for pulling out a single signed page, one chart, or a single invoice from a much longer document.
Yes — that's the most common use case. Click any combination of thumbnails (pages 5, 12, and 47 all in the same selection), or type "5, 12, 47" in range mode. The output PDF contains exactly those pages, in document order, with nothing in between.
Yes. the page-copy step moves the source page objects byte-for-byte into the new PDF — no rasterisation, no re-encoding, no compression. Embedded fonts, images, and vector graphics arrive in the output identical to how they were in the source.
Internal hyperlinks and form fields are preserved when their target stays within the selection — a link from page 3 to page 7 survives if both pages are kept, and breaks (becomes a no-op) if page 7 was excluded. Bookmarks are not currently carried into the extracted PDF; if your workflow depends on bookmarks, leave a note and we'll prioritise that update.
Roughly proportional to the page count, with some overhead from shared resources. Extracting 10 pages from a 100-page PDF typically gives you a file around 10-15% of the source size — the page content drops linearly, but shared font subsets and image objects don't deduplicate quite as efficiently as you'd hope. The result is still much smaller than the original.
Not directly — the file needs to be unlocked before extraction. When you drop an encrypted PDF in, the tool detects it and asks you to unlock it. Our Unlock PDF tool removes owner restrictions automatically and decrypts password-protected files locally with the password you supply — extract from the unlocked copy.
Yes. The thumbnail grid is responsive — three columns on phone, four on small tablets, more on desktop. Tap any thumbnail to toggle selection. For very long documents (200+ pages), the range input is usually faster than scrolling through hundreds of thumbnails on a small screen, so the mode toggle is right there to switch to.
Because that would just re-save the same PDF — a copy with the same number of pages and identical content. That's almost never what someone means by "extract pages". The button is disabled until at least one page is deselected, to keep you from accidentally shipping a no-op rename. If you genuinely need a copy of the original, use your operating system's file-copy.
No — extracted pages always come out in original document order, regardless of the order you clicked them. This matches what most people expect ("I asked for pages 5, 2, and 8 — give me a PDF with those pages") and is the safest default. If you do need a different order, run the extracted PDF through our Merge tool after, which lets you reorder pages with drag-and-drop.
Not in a single batch — this tool processes one PDF at a time, since cherry-picking is intrinsically per-document (page 5 means something different in each file). To work across multiple PDFs, run them one after another and combine the resulting extractions with our Merge tool if you want a single output file at the end.
Tools picked because they pair naturally with the one above — the next step in a typical PDF workflow.