E-book formats explained: EPUB, PDF, Kindle and more
An e-book is not one kind of file. Here is what is inside the common formats, why they differ, and what that means for readers, authors and anyone who needs a book read aloud.
When people say "e-book" they usually picture the text of a novel on a phone or an e-reader. Behind that picture sit several file formats with different histories and different strengths. Knowing the difference is useful whether you are a reader trying to move your library to a new device, an author publishing your own book, or someone who relies on a screen reader and needs a file that actually works with it.
Reflowable and fixed layout: the basic split
Every e-book format falls on one side of a basic divide.
A reflowable e-book has no fixed pages. The text flows to fit whatever screen it is shown on, and the reader can change the font, the type size, the line spacing and often the colours. A page on a phone holds fewer words than a page on a tablet, and the "page numbers" are calculated by the reading software rather than fixed. This is ideal for prose, where the words matter and the exact position of a line break does not. Most novels and much non-fiction are published this way.
A fixed-layout e-book keeps the design of each page exactly as it was made, like a printed page. That suits picture books, comics, cookbooks, textbooks with complex layouts and art books, where the relationship between text and image is the point. The cost is flexibility: on a small screen the reader may have to zoom and scroll, and enlarging the type is harder or impossible.
Much of the frustration people feel with e-books comes from a book being in the wrong kind of layout for the device or the reader. A cookbook converted to reflowable text loses its design; a novel locked into fixed pages becomes tiring to read on a phone.
EPUB: the open standard
EPUB is the open, widely supported format for e-books. Under the surface an EPUB file is a ZIP archive containing web technologies: the text is written in XHTML, styled with CSS, and accompanied by a package file that lists the contents, the reading order and the metadata (title, author, language, identifiers and so on). Because it is built from the same ingredients as web pages, EPUB can handle reflowable text well, and EPUB 3 added support for fixed layouts, embedded audio and video, scripted interactivity, mathematical notation and text in many writing systems, including vertical and right-to-left scripts.
The standard was developed by the International Digital Publishing Forum and is now maintained by the World Wide Web Consortium (W3C). The current version, EPUB 3.3, became a W3C Recommendation in May 2023, alongside a companion standard on accessibility.
EPUB is the format used by many e-book stores and reading apps, and by most library lending platforms. If you buy an e-book outside Amazon's store, it is likely to be an EPUB, though it may be wrapped in digital rights management that ties it to a particular app or account.
Kindle formats
Amazon's Kindle ecosystem has used its own formats. The early Kindle format grew out of Mobipocket, a format Amazon acquired before launching its first reader in 2007. Later formats added richer layout and typography. Books bought from the Kindle store are delivered in Amazon's formats and usually carry Amazon's DRM, which is why they open only in Kindle devices and apps.
For authors the situation has become simpler. Amazon's self-publishing service accepts EPUB files for upload and converts them itself, so an author can produce one well-made EPUB and use it for Amazon and other stores alike. Amazon's Send to Kindle service also accepts EPUB files from readers for personal documents.
The practical consequence for readers is that the format matters less than the store. A Kindle book lives inside Amazon's system; an EPUB from another store lives inside that store's app or a compatible reader. Moving between them is generally possible only for books sold without DRM.
PDF: a printed page on a screen
PDF was designed to reproduce a printed page faithfully on any computer, and it does that job very well. That makes it the natural format for documents that must look exactly the same everywhere: forms, academic papers, manuals and books typeset for print. But it is a fixed-layout format by design, and on a small screen a book-sized PDF page is hard to read. PDFs can be made accessible, with tags that describe the structure and reading order, but many are not, and an untagged PDF can be nearly unusable with a screen reader.
Many technical and academic publishers still sell PDFs alongside EPUBs, because their readers want to cite page numbers that match the printed edition.
Other formats you may meet
- Plain text and HTML. Many public-domain books are distributed as simple text or web pages, which work everywhere but carry little formatting. Projects such as Project Gutenberg and Standard Ebooks offer public-domain titles in several formats, including EPUB. Our article on the public domain explains which books are free to copy.
- Comic book archives (CBZ, CBR) are simply collections of page images in an archive, used for comics and manga.
- DAISY is a format designed for accessible reading, combining structured text with synchronised audio. It was developed by the DAISY Consortium, which also contributed heavily to accessibility in EPUB 3.
- Audiobook files are a separate world, covered in our article on audiobooks.
Accessibility: why the file structure matters
For readers who are blind, partially sighted or dyslexic, or who have difficulty holding a book, a well-made e-book can be transformative. Text can be enlarged, re-coloured, read aloud by text-to-speech, or sent to a refreshable braille display. But all of that depends on how the file is built. A book needs proper headings, a logical reading order, text descriptions for meaningful images, correctly marked languages and real text rather than pictures of text.
The W3C's EPUB Accessibility standard sets out what an accessible EPUB should contain and how a publisher can declare a book's accessibility features in its metadata, so that a reader can find out before buying whether a book will work for them.
Regulation has pushed in the same direction. In the European Union, the European Accessibility Act (Directive (EU) 2019/882) covers e-books, e-reading software and the online sale of e-books, and its requirements have applied since 28 June 2025. Publishers selling into the EU have had to review how their e-books are produced and described as a result. Internationally, the WIPO Marrakesh Treaty allows accessible copies of books to be made and shared across borders for people with print disabilities.
Which format should you use?
If you are a reader
Choose your store and device together, because that decision locks in more than the file format does. If you want to be able to move your books between devices and apps, look for titles sold without DRM, and keep copies of the files. Our guide to e-readers compares dedicated readers, tablets and phones.
If you are an author or small publisher
Make a clean, valid, accessible EPUB 3 file and treat it as your master. It can be uploaded to most retailers and library distributors, and converting from a good EPUB is far easier than converting from a word-processor file. Keep a print-ready PDF for print on demand, and consider fixed layout only for books whose design genuinely requires it.
Sources
- W3C, EPUB 3.3 specification, and press release on EPUB 3.3 becoming a W3C Recommendation (May 2023).
- EUR-Lex, Directive (EU) 2019/882, the European Accessibility Act.
- WIPO, the Marrakesh Treaty.
- Amazon Kindle Direct Publishing, KDP help pages on supported manuscript formats.
- DAISY Consortium.
- Wikipedia, Mobipocket (Amazon's acquisition and the early Kindle format).