Clean up scanned book pages and split them in half
Scanned a book or magazine two pages at a time? Make the faded pages easier to read, and turn every double-page spread into two single pages, in reading order, for a phone, tablet or e-reader. The fold is found on each page automatically, even when the spread is off-centre, and you can drag the cut where you want it.
By default the scan is not re-compressed, so the pages look exactly as sharp as before and the file stays the same size. Want darker text on scanned pages? Use the second download button. Free, and your PDF never leaves your browser.
See it work on a sample book (12 s)
Choose a file
Pages
Drag the line to move a cut. Click a page to split it or not. Shaded borders are trimmed.
Darker scans make faded text easier to read, but can change the colours of photos and pictures. Use Download single pages to keep them as they are.
Everything happens in your browser: the PDF is read, analysed and rebuilt on your device. Nothing is uploaded.
How to split PDF pages in half
- Choose the PDFOr JPG/PNG photos of the spreads, which become a PDF first.
- The folds are foundEach page is looked at on its own: a wide page with a fold, where two pages meet, is a two-page spread, and the cut goes on the fold, not just down the middle.
- A text model double-checksA small text-finding model runs in your browser and checks each page: text on both sides of the fold means two pages. Only pages where the two checks disagree are outlined for you to look at.
- Check and adjustDrag the line on any page to move its cut, or click a page to split it or leave it whole. Covers, fold-outs and single pages stay as they are, even wide ones.
- Pick the orderLeft page first, or right page first for Arabic, Urdu, Hebrew and Japanese books and manga.
- DownloadEvery spread becomes two pages in the new PDF, in reading order.
Why split a two-page scan?
| Read on a phone or e-reader | One book page fills the screen instead of two tiny ones. |
|---|---|
| Page numbers match | Page 41 of the PDF is page 41 of the book again, which makes citations and bookmarks work. |
| Better OCR and search | Text recognition reads one column of text at a time instead of two side by side. |
| Printing | Single pages print the right way round on A4 or Letter paper. |
| Easier book reading | Each page is one full book page at a readable size, so scanned books, magazines and textbooks are easier to read on any screen. |
Questions
How do I split a scanned PDF page into two pages?
Open the PDF here. Every page that is wider than it is tall and has a fold down it is treated as a two-page spread and split where the pages meet. Check the line on each page, move it if needed, and download. The new PDF has one book page per PDF page.
Will the quality of my scan get worse?
“Download single pages” does not re-compress or redraw anything: each spread becomes two PDF pages that show different halves of the same image, so they are as sharp as the original and the file is about the same size. “Download with darker scans” redraws only the scanned pages (pages with no text): they are rendered at about 144 dpi on A4 or Letter and saved as JPEG at quality 90%, so fine detail can be lost. Pages with text are not touched. It can also change the colours of photos and pictures, so for those use the first button, which keeps everything as it is.
What if the spread is not centred?
The cut is not simply the middle of the page. The tool looks for the fold itself: the narrow band without text where the two pages meet, often lighter or shadowed. You can also drag the line on any page.
What happens to a wide cover or a fold-out map?
It stays one page. A wide page is only split by default when a real fold is found: near the middle, with text or pictures running down both sides of it. A wide cover, fold-out or landscape chart has no such fold, so it is kept whole and marked “wide, no fold”. Click it if you do want it split.
Can I split only some pages?
Yes. Covers, title pages and anything already single are left whole by default. Click a page to change it, or choose “every page” or “no pages” and then pick.
Do I have to check every page of a long book?
No. Each page is checked twice: once for the fold, and once by a small text-finding model (PaddleOCR) that runs in your browser and sees whether there is text on both sides of it. When both agree, the page is settled. Only pages where they disagree are outlined, and even those follow what the rest of the book does. The model downloads once (about 15 MB) and checks about two pages a second on a laptop.
What does “Trim borders” do?
It takes off what is around the paper: the black scanner lid, a white or grey background, the table under a photographed book. Only flat borders and a clear paper edge are trimmed, never into text, and the trimmed parts are shaded on each page before you download. It is a crop, so nothing is re-compressed. Untick it to keep the pages as they are.
Does it work with photos of a book?
Yes, choose JPG or PNG photos; they become a PDF, one photo per page, and are split the same way. Pages photographed at a strong angle or with a curved spine are cut straight down; flattening curved pages is not part of this tool yet.
Does splitting make a scanned book easier to read?
Yes. Each spread is turned into two full pages, so the text is larger on a phone, tablet or e-reader, and you can read straight through the book without zooming and panning across two pages at once. The scan itself is not changed.
Is my PDF uploaded?
No. The PDF is read, analysed and rebuilt in your browser. Nothing is sent anywhere, so it works for private documents too.
Can I make a faded scan easier to read?
Yes, for scanned pages. Use “Download with darker scans”: the ink gets darker and the paper whiter, one colour channel at a time, so coloured headings and diagrams keep their colour. Pages that already have selectable text are left as they are. Scanned pages were already images with no selectable text, so this does not remove any; run OCR PDF afterwards if you need searchable text.
Does it work with right-to-left books?
Yes. Choose “right page first” for Arabic, Urdu, Hebrew and Japanese books and manga, so the pages come out in reading order.