Split a PDF — without uploading it
Extracting pages from a 200-page report or separating an invoice from a contract — everyday tasks. The split here runs locally with pdf-lib + PDF.js, so nothing leaves your machine.
- PDF never touches a server
- Range syntax: '1-3, 5, 7-10'
- 100 pages split in ~1 second
- Outputs auto-bundle into ZIP
Pick pages. Get clean splits.
Split loads your PDF, shows thumbnails, and lets you pick which pages go where. All in your browser.
Open the Split tab
Click the button above or go to /split.
Drop your PDF
Thumbnails render via PDF.js — 100 pages takes 2–5 seconds.
Pick a split mode
Click pages to include, type a range like '1-3, 5, 7-10', or split every page into its own PDF.
Choose output format
Bundle outputs into a ZIP (default) or download them individually.
Hit Split & Download
pdf-lib's copyPages does the work. 100 pages split 5 ways takes ~1 second.
What you can expect from Split.
Practical notes.
Input format
PDF only. For Word/Excel/image, convert to PDF first with Combine.
Output quality
Bookmarks, links, annotations, fonts — all copied via pdf-lib's copyPages. No re-encoding.
Range strings
Comma-separated. '1-3, 5, 7-10' gives three PDFs. Single page numbers valid too.
File size
No enforced cap. Desktop Chrome handles 500 MB+; phones around 50–100 MB.
Sensitive docs stay yours.
Splitting often involves sensitive documents — contracts, financial statements, medical records. Exactly the documents you don't want on someone else's server.
- Sejda and SmallPDF upload the source PDF just to render thumbnails
- "Delete after 2 hours" still means your file sat on their disk
- Free tier caps splits at 3/day or 200 pages
- Thumbnails render locally via PDF.js — no upload
- Ranges parsed in your tab, not on a server
- No 200-page cap; only RAM matters
Browser-side splitting has real limits.
Where you'd want a different approach.
Batch automation
No API. For pipelines, use pdftk or qpdf.
Very large PDFs (1GB+)
Browser memory struggles. Server tools handle these better.
Splitting by content
Splits by page number, not content. For semantic splitting, OCR + script.
Removing pages, not splitting
If you just delete pages and keep one output, that's Edit — not Split.
Questions people ask about splitting.
Can I split a password-protected PDF?
Not directly. Unlock first with the password or qpdf, then split.
Difference between 'Pick pages' and 'By range'?
'Pick pages' uses thumbnails — contiguous selections become one output, gaps split. 'By range' lets you type the same thing as a string.
Does the output keep original quality?
Yes — pages are copied bit-for-bit. Images, fonts, vectors preserved.
My PDF is 800 MB. Will this work?
Depends. Desktop Chrome with 8GB+ RAM usually handles up to 1 GB. Mobile is risky over 100 MB.