EPUB Validator

An EPUB is a ZIP archive with a fixed layout inside. The tool reads it the way a reading app does and checks each layer.

The package

• The mimetype file: present, first in the archive, uncompressed and containing exactly application/epub+zip.
• META-INF/container.xml: present, readable, and pointing to a package file (.opf) that exists.
• A damaged archive: a missing file index or a book cut short during a download. Everything up to the cut is kept.

Metadata

• A title, a language and an identifier, and a unique-identifier that points to that identifier.
• For EPUB 3, a last-modified date in the required format.

When one is missing, the tool adds a sensible value: the first heading as the title, en as the language, a new identifier. Check these in the fixed book and change them if they're wrong.

Manifest, reading order and table of contents

• Files listed in the manifest that aren't in the book, and files in the book that the manifest leaves out.
• Wrong or missing media types, including old font types such as application/x-font-ttf.
• Duplicate ids and files listed twice.
• A reading order that points to unknown items, or is empty.
• For EPUB 3, a missing navigation document. The tool builds one from the reading order.

Chapters

• XHTML that isn't well-formed: tags left open, a stray &, characters XML forbids.
• HTML habits XHTML doesn't allow, such as   or a line break tag that isn't closed.
• Images and stylesheets the chapter links to but the book doesn't contain.

Protected books

Books with DRM are reported and left alone. We don't remove DRM.

If it doesn't work

• The tool says the book has no readable chapters: the content files are missing or unreadable. Export the book again from the app you wrote it in.
• The fixed book opens, but the title is "Untitled": the original had no title. Set it in your writing app, or edit the dc:title line in the package file.

EPUBCheck is the official EPUB conformance checker, maintained under the W3C. It runs on Java and checks the whole specification. This tool is not EPUBCheck: it runs its own checks in your browser, covers the structural errors that most often stop a book from opening or uploading, and fixes them instead of only listing them.

Matching EPUBCheck messages

If EPUBCheck or a store gave you an error code, these are the ones the tool fixes:

• PKG-006: the mimetype file is missing or not first.
• PKG-007: the mimetype file has the wrong content.
• RSC-002: META-INF/container.xml is missing.
• OPF-002: the package file isn't where container.xml says.
• RSC-005 and RSC-016: a file couldn't be parsed, usually broken XML or a missing required element.
• RSC-007 and RSC-001: a referenced file isn't in the book.
• OPF-003: a file is in the book but not in the manifest.
• OPF-013: a file's media type doesn't match its content.
• OPF-030: the unique-identifier points to no identifier.
• OPF-049: the reading order points to an unknown item.
• OPF-074: one file is listed in the manifest more than once.

What it doesn't check

CSS, accessibility, fixed-layout rules, scripts, links between chapters, and EPUB 2 books without an NCX table of contents. Before you publish, run the fixed book through EPUBCheck or the store's own previewer as a final check.

1. Open the fixed .epub in a reading app and page through it.
2. Check the table of contents. If the tool built one, it lists the chapters in reading order with their first heading as the title.
3. Check the title, author and language the store will show.
4. Upload it, or send it to your reading device.

If a store still refuses the book, see kindle rejecting epub for Kindle and KDP, or run EPUBCheck to see the remaining errors.