EPUB Checker

What it's for

An EPUB is a ZIP of web pages with strict rules about its packaging, and stores and readers (Apple Books, Kobo, Google Play Books, Send to Kindle) reject or misshow books that break them. This checks a book for the common mistakes and shows its details, cover and table of contents.

For example, when Send to Kindle rejects a book you made with a script or by editing the files, check it here: a mimetype file that isn’t first in the ZIP, or a picture the text points to that isn’t in the book, are typical causes, and both are named.

How to use it

Choose an EPUB (or drop it on the page). You see its cover, then its details (title, author, language, identifier, publisher, dates, EPUB version, the number of parts in its reading order, an approximate word count and the number of files), then the problems found, worst first:

• Packaging: mimetype must be the first file, stored uncompressed, containing exactly application/epub+zip; META-INF/container.xml must point to the package file.
• Metadata: a title, language and identifier, and for EPUB 3 a dcterms:modified date; accessibility metadata, which the European Accessibility Act expects.
• Manifest and spine: files that are listed but missing, files present but not listed, wrong media types, duplicate ids, reading order entries that don’t exist or aren’t pages, and the navigation document (EPUB 3) or NCX (EPUB 2).
• Pages: XHTML that isn’t well-formed (such as  , which XHTML doesn’t know), links and pictures that point to missing files, pictures without alt text, and scripts.

The Table of contents is listed as the book gives it.

Example

Make a copy of an EPUB, unzip it, delete one picture from its images folder and zip it again with your usual zip tool. Checking it gives an error that the deleted picture is listed in the manifest but isn’t in the book, another for each page that shows it, and very likely one that mimetype isn’t first or is compressed.

Good to know

These are the common problems, not all of the official EPUBCheck’s hundreds of rules: CSS, fonts, fixed-layout details and media overlays aren’t checked, and links to a place in a page (#id) are checked only for the file. DRM-protected books can’t be read. The book stays in your browser.