Split PDF — Extract Pages Online
Split a PDF into separate files by page range.
Drop your PDF file hereTap to select a file
Supports PDF, up to 200MB
What to do next
Related tools
About Split PDF
Splitting a PDF is what you do when a file came to you bigger than the part you actually need. Maybe a 320-page medical-records release contains the four pages your insurer asked for. Maybe a year-end financial statement bundles 12 months of reports into one document and you only need Q3. Maybe a PDF arrived as a scanned 50-page contract and you want to share each signed page separately. Favtoo's Split PDF tool covers all three: extract a single range, split into many ranges at once, or break the file down to one PDF per page — without ever uploading anything to a server.
The split operation uses pdf-lib to copy the requested page objects into freshly-built output PDFs, which means three useful things happen automatically. First, the visual fidelity is exact: every output page is byte-identical to the source page, including embedded fonts, vector graphics, and image streams. Second, the file sizes scale linearly with the page counts, so an extracted 5-page section from a 50-page document is roughly one tenth the size of the original. Third, the operation is fast — there is no rasterisation step, so even a 500-page document splits in seconds.
The interface is range-driven. Type ranges using the standard PDF range syntax: 1-3 for the first three pages, 5 for a single page, 7-10, 12, 14-end for a more complex selection, and the tool parses, validates, and previews exactly what you will get. If you prefer to split visually, the page thumbnails on the left let you click-select pages and see your selection update in real time. The output is always a ZIP file when you produce more than one PDF, and a single PDF when you do not — no browsing through folders to find the result.
A practical note about page-numbering: when you extract pages 7–10 from a long document, the resulting 4-page PDF is renumbered 1–4 in the output. The page tree gets rebuilt from scratch, so anything that referenced an absolute page number in the original (a hyperlink, a TOC entry, an outline bookmark) is rewritten to point at the new sequence. This is correct behaviour — if you need the original page numbers visible on the page, run the result through Add Page Numbers to PDF with a custom starting number.
How it works
- 1Upload the PDF you want to split. Page thumbnails render in the sidebar so you can see what is in the file.
- 2Type the page ranges you want, using syntax like 1-3, 5, 7-10. Or click pages directly to build the selection visually.
- 3Pick the split mode — single output, separate output per range, or one PDF per page.
- 4Hit Split. pdf-lib copies the requested page objects into one or more new PDFs in a single pass.
- 5Download the result as a single PDF or, when there is more than one output, a ZIP archive of named PDFs.
Common use cases
- Pull just the four signed pages out of a long contract to send to a counterparty
- Extract one chapter from a 400-page reference PDF for offline reading
- Break a multi-invoice export into one PDF per invoice for your accounting tool
- Split a scanned passport-and-visa scan into two separate ID images
- Separate the appendices from the body of a research paper for selective sharing
- Slice a 12-month statement PDF into 12 monthly files for archive storage
FAQ
Can I split by page ranges?
Yes — enter ranges like 1-3, 5, 7-10 to extract exactly the pages you need.
Does splitting change the content?
No — each output file contains the original pages with no quality loss.
What if my PDF is password-protected?
Remove the password first with our Remove PDF Password tool, then split.
What page-range syntax does the tool understand?
You can use any combination of single numbers, hyphenated ranges, and the keyword "end". Examples: 1-3 (first three pages), 5 (just page 5), 7-10, 12, 14-end (mixed). The parser is whitespace-tolerant and ignores duplicates so 1-3, 2 still produces pages 1 to 3, not page 2 twice.
Will the split files keep the same quality as the original?
Yes — exactly the same. The split operation copies page objects without re-rendering, so embedded fonts, vector artwork, image resolution, and OCR text layers are byte-identical to the source. There is no quality loss whatsoever.
Do I get one ZIP file or many separate downloads?
When you split into one output, you get a single PDF. When you split into multiple ranges or one-per-page, the tool packages them into a ZIP so your browser does not have to fire dozens of download dialogs. Each PDF inside the ZIP is named based on the source filename and the range it contains.
How are pages renumbered in the output PDFs?
Each output PDF starts at page 1 internally, regardless of the page numbers in the source document. If you need the visible page numbers on the page to reflect the original sequence, run the result through Add Page Numbers to PDF and set a custom starting number.
Can I split a password-protected PDF?
You will need to remove the password first using Remove PDF Password (you must know the correct password). Once the file is unlocked, splitting works as normal. You can re-protect any of the resulting files with Protect PDF afterwards.
Are bookmarks, annotations, and form fields preserved in each output?
Bookmarks pointing at pages within the extracted range are preserved and re-targeted to the new page numbers. Annotations (highlights, sticky notes, ink markup) on extracted pages carry through. AcroForm field values are preserved for any fields that live on the extracted pages.
How large a PDF can I split?
The 200 MB upload limit applies. In practice the bottleneck is your device’s memory rather than the tool — splitting a 500-page text-heavy PDF takes only a few seconds; splitting a heavily image-laden 180 MB PDF may take 10–20 seconds on a mid-range laptop.
What happens if I enter a page range that is out of bounds?
The parser flags it before the split runs, so a request like 1-999 on a 50-page PDF is caught immediately with a friendly error rather than producing a half-broken output. Fix the range and re-run.