PDF to EPUB Converter — Free, Online, No Upload

A PDF keeps every page exactly as it would print. On a 6-inch e-reader or a phone, a full Letter or A4 page shrinks until the text is hard to read, so you zoom in and scroll sideways on every page. An EPUB has no fixed pages. The reading app lays the text out to fit the screen, and you pick the font size, margins and line spacing.

This converter reads the text inside your PDF and rebuilds it as a book you can resize. It finds where each paragraph starts from the space between lines. Pages with two columns are read one column at a time, using the same engine as our PDF to Text converter. Chapter titles become real headings with a table of contents, photos are kept, and page 1 becomes the cover. It all runs in your browser, so the PDF is not uploaded anywhere.

Kobo e-readers and Apple Books (iPhone, iPad and Mac) open EPUB files directly, and you can upload them to Google Play Books. A Kindle can't open an EPUB copied over USB, but Amazon's Send to Kindle service accepts EPUB files and adds them to your Kindle library.

Going the other way, for printing or a fixed layout? Use EPUB to PDF. Need the text to edit rather than to read? PDF to Word gives you a .docx file.

You get the words, in reading order. We tested a 10-page book PDF with 3 chapters, a 120-page report with 40 parts, and one page of a two-column research paper. Every paragraph came out in order, with the lines of each paragraph joined so they reflow. A paragraph that runs onto the next page is joined back into one. All 880 paragraphs of the 120-page report came through. Each file took about 2 seconds.

Chapters, headings and a table of contents. If the PDF has bookmarks, they become the book's table of contents, and each chapter starts on a new page. Our 10-page book got 9 entries: 3 chapters with 2 sections each. Without bookmarks, titles set in bigger type than the body text become the headings. The same book without bookmarks got the same 3 chapters and 6 sections, plus its title page. Headings that are only bold, at the same size as the text, are not found.

Photos come along. Pictures in the PDF are placed in the EPUB where they sit on the page, shrunk to at most 1,600 pixels on the long side. Photos are saved as JPG, and pictures with see-through parts as PNG. A small logo that repeats on many pages is left out, and so are full-page backgrounds. Charts drawn with lines and shapes, not stored as pictures, are not kept; their labels come through as short lines of text. Page 1 becomes the cover image.

Running headers and page numbers are removed when they repeat at the top or bottom of the pages. Fonts, bold and italic, colours, links and tables are not kept.

Scanned PDFs need OCR first. A scan is a set of pictures of pages, with no text inside. Instead of an empty book, you get a message that the file needs OCR (text recognition).

The book is a standard EPUB 3 file. It also has an older-style table of contents, so older e-readers can use it. Every part of it passed a strict XML check in our tests. EPUB output has no watermark on the free tier. Free use is 2 files a day, up to 10 MB each; Pro takes PDFs up to 500 MB.