Scans become real text
An OCR model reads every page, then we check it against any text the PDF already has and put back what it missed.
Elsewhere: no text layer, no text. Scans come out as pictures of pages, or blank.
PDF to EPUB, for people who read
ocrepublic reads scanned novels, two column research papers and other stubborn PDFs the way you do, and hands back an EPUB for Kobo, Kindle and Apple Books. Chapters land in the contents and footnotes link to their notes. Page numbers stay out of your sentences.
Convert 10 pages freeNo card needed. You see the exact price before you pay.
100% of the books on our test shelf pass epubcheck, the official EPUB validator, with zero errors and zero broken links.
Free converters copy the page as it lies, clutter and all. We marked the four things they carry into your book, then fixed each one.
The village had no road worth the name until the spring of 1887, when the council agreed at last to pay for gravel from the quarry at 144 Harrow End. The Long Valley 1171 The work took two summers.
Most of the men who laid it were the same farmers who would use it, and they argued over every2
14 The quarry closed in 1931; its ledgers survive in the county archive.3yard. One of them kept a diary, and from it we know the price of a cart of stone.
The village had no road worth the name until the spring of 1887, when the council agreed at last to pay for gravel from the quarry at Harrow End.14 The work took two summers.
Most of the men who laid it were the same farmers who would use it, and they argued over every yard. One of them kept a diary, and from it we know the price of a cart of stone.
14. The quarry closed in 1931; its ledgers survive in the county archive. Back to text
A passage written to show the typical faults side by side. Real, unedited output from a free converter is further down the page.
Free conversion tools copy the text of a PDF in the order it happens to be stored. They were never built to read a page the way you do.
An OCR model reads every page, then we check it against any text the PDF already has and put back what it missed.
Elsewhere: no text layer, no text. Scans come out as pictures of pages, or blank.
Running heads and page numbers are removed, even the kind that changes on every page and defeats find and replace.
Elsewhere: they turn up mid sentence every screen or two.
Notes can move to a Notes section at the end, each with a link back. 516 of 529 linked on our hardest test book.
Elsewhere: notes land wherever the printed page happened to end.
Paragraphs that run over a page break are joined into one. Two column research papers are read left column first, then right, as you would.
Elsewhere: paragraphs split at every page, lines from both columns mixed together.
Chapters appear in your reader's table of contents, the PDF's own links keep working and figures are embedded.
Elsewhere: often one long file with no chapters to jump to.
Journal articles, conference papers, textbooks, old two column editions. We read the layout first, then the words, so the text comes out in the order the author meant.
Left column first, then right, joined into one column that reflows on any screen.
Each figure is cropped from the page and embedded beside its caption, not lost.
Tables and simple formulas are kept. Pages that need a second look are flagged in your summary.
Bibliography entries stay whole, and footnotes link to a Notes section with a way back.
A PDF is a picture of a page.
We give you back the book.
A typical free converter against ocrepublic. Run the free preview and check these on your own book.
| What you want | A free converter | ocrepublic |
|---|---|---|
| Scanned books turned into text | Rarely | Every page |
| Two column research papers read in order | Columns often mixed | Left, then right |
| Running heads and page numbers removed | No | Automatic |
| Footnotes linked at the end | No | With links back |
| Paragraphs joined across pages | Hit and miss | Yes |
| Tables, formulas and figures kept | Often flattened or lost | Kept, figures embedded |
| Every page checked twice | No | Against the PDF's text |
| See the result before paying | Free | 10 pages free |
| Your time | Minutes, then fixing by hand | About 2 minutes, ready to read |
Paradise Lost, Book I, PDF page 12: plain verse, no footnotes, no running heads. A best case for any converter, and the free one still joined three pairs of lines.

The mother of mankind, what time his pride
Had cast him out from Heaven, with all his host Of rebel Angels, by whose aid, aspiring
To set himself in glory above his peers,
He trusted to have equalled the Most High,
If he opposed, and with ambitious aim
Against the throne and monarchy of God,
Raised impious war in Heaven and battle proud, With vain attempt. Him the Almighty Power
Hurled headlong flaming from th’ ethereal sky, With hideous ruin and combustion, down
Output copied as it came, nothing edited.
The mother of mankind, what time his pride
Had cast him out from Heaven, with all his host
Of rebel Angels, by whose aid, aspiring
To set himself in glory above his peers,
He trusted to have equalled the Most High,
If he opposed, and with ambitious aim
Against the throne and monarchy of God,
Raised impious war in Heaven and battle proud,
With vain attempt. Him the Almighty Power
Hurled headlong flaming from th' ethereal sky,
With hideous ruin and combustion, down
Every line where Milton put it.
Drop a PDF of up to 1,000 pages. You see the page count and the exact price first. Blank pages are free.
The first 10 pages convert free. Compare them with your PDF, side by side.
Convert the whole book and download the EPUB. Pages we could not read go back to your balance.
“Reading papers on my commute is now possible.”
“Scanned and preserved some books I can never find online. Thank you so much!”
“Really handy tool, simple and easy to use.”
One credit converts one page. No subscription, and credits never expire.
of 300 pages each
Choose smallof 300 pages each
Choose mediumof 300 pages each
Choose largeFirst 10 pages free. Judge the result on your own book, no card needed.
Blank pages cost nothing. We find them and leave them off the bill.
Unreadable pages refunded. Their credits return to your balance automatically.
Credit packs themselves are not refundable once delivered, which is why the preview is free.
PDFs up to 1,000 pages and 100 MB, text or scanned. Encrypted or password protected PDFs cannot be converted.
Yes. Two column research papers are read in order, left column then right, and come out as one column that reflows on your screen. Figures are embedded with their captions, tables and simple formulas are kept, and bibliography entries stay together. Footnotes can move to a linked Notes section at the end. Pages with unusual layouts are checked page by page and flagged in your quality summary if they need a look.
The reading and cleanup rules are built and tested for English and German, including German compounds split across line ends. Other languages are untested here, so the honest answer is to run the free 10 page preview and judge it on your own book.
Anything that opens EPUB 3: Kobo, PocketBook, Apple Books and most Android reading apps. For Kindle, use Send to Kindle, which accepts EPUB.
A few minutes for a typical book, closer to ten for a 1,000 pager. The progress page ticks through each page and stage, and you can close the tab: the job keeps running, and the result waits in your history.
Only you, and we keep nothing. Your PDF is deleted as soon as the conversion finishes, and the EPUB one hour after your first download, or after 24 hours if you never download it. Each page image is read by our OCR provider, Z.ai, which processes it outside the EU. Your files are used for nothing else.
Their credits go back to your balance automatically, and your quality summary names the pages.
For one hour after your first download, in case it did not finish. After that the file is gone, because we keep nothing, and converting the book again costs credits. Save it somewhere safe.
Yes, but there is no password. Sign in with a code we email you, or with Google or Microsoft.
Every page is read by an OCR model, and each page costs us money to read, so each page costs you a credit. Free converters skip that step and copy whatever text the PDF already has, which is exactly where they fail on scans.
Calibre is excellent free software, and the best known free option. It manages an ebook library beautifully and converts simple text PDFs well. Its PDF converter reads the same text layer that other free tools read, so it meets the same limits: scans need OCR it doesn't do, running heads and page numbers stay in the text unless you write removal rules, footnotes land mid paragraph, and two column pages can come out mixed. Pandoc is another excellent free tool, built for converting between document formats, but it does not read PDFs at all: you would first have to turn the PDF into text or Markdown with another tool, and the same limits come along. If your PDF is a plain novel, Calibre may be all you need. For anything harder, preview 10 pages here and compare.
Send us its first 10 pages. They're on the house.
Convert 10 pages freeNo card needed. You see the exact price before you pay.