OCR included

Scanned PDF to EPUB converter, with OCR

To convert a scanned PDF to EPUB, upload it here: epubwise reads pages that have no real text with OCR, repairs scrambled scan text, and rebuilds the paragraphs.

Convert a PDF to EPUB

No sign-up
Deleted within an hour
Reads scanned PDFs

Free: 10 files a day, OCR on up to 100 scanned pages a day (a pass reads up to 300 per file). Only convert files you have the right to convert. Password-protected and DRM-locked PDFs can't be converted.

tom_sawyer_1876.epub epubwise

he saw a new girlTOMSAWYER.inthegardenalovely

he saw a new girl in the garden a lovely little blue-eyed
creature with yellow hair plaited into two long tails,
white summer frock and embroidered pantalettes.

(7 pictures of pages. No text.)

all the summer world was bright and fresh, and brimming
with life. There was a song in every heart ;. and. if the
heart was young the music issued at the lips. There was in
every face and a spring in every step.

The 1876 first edition of Tom Sawyer, scanned, and chapter II with its text layer removed. PDF24 came closest of the tools we tested. On the chapter with no text layer it kept pictures only; epubwise read 97.6% of the chapter's words with OCR. Unedited output.

Why scans usually convert badly

A scanned PDF is a photo of each page. Some scans carry an invisible text layer made by OCR software, and many of those layers store words out of order, in pieces, with a different size for every line. A converter that trusts the layer gives you scrambled sentences, a new paragraph on every line, and stray numbers in the middle of the story.

Older books add their own trouble: folded maps, statistical tables and diagrams whose labels are scattered across the page as dozens of tiny fragments.

What happens to a scan here

OCR only where it's needed. Each page is checked, and only pages without usable text are sent through OCR.
Scrambled layers put back in order. Fragments of each line are rejoined left to right, and uneven letter sizes are evened out so paragraphs stay whole.
Dust and specks dropped. Stray dots and bars picked up by the scanner don't turn into one-character paragraphs.
Maps and charts saved as pictures. That part of the page is kept as an image, so its labels and figures stay where they belong.
Headers taken out. The book title and folio printed at the top of every scanned page are removed from the text.

On the free plan, OCR runs on up to 100 scanned pages a day; the result tells you if any pages were left unread. See what the free plan covers, or read about each conversion step in detail.

After the conversion

You get a standard EPUB that opens in any reading app. Old scanned books are often easier to read this way than in the original PDF, because the text can be enlarged without scrolling sideways. To read it on a Kindle, send the EPUB through Amazon's service. For a PDF that was made on a computer rather than a scanner, the standard PDF to EPUB converter works the same way.

Questions about scans

How long does a scanned book take?

A scan that already has a text layer converts in a minute or two. Pages that need OCR take a few seconds each.

Will the OCR be perfect?

No OCR is. Clean, straight scans come out very well; faint, skewed or handwritten pages produce more errors.

Can I convert photos of pages?

Put the photos into a PDF first (most phones can print to PDF), then convert the PDF.