Inspect
What is this file, and why won’t it open?
Identify a book from its contents rather than its name, read its metadata and structure, and — for an EPUB — get a report of the faults that stop readers opening it, each with what to do about it.
No upload· Files are read in this tab and never sent anywhere
Intake
Drop a file to inspect
Any book file. It is identified from its contents, so a renamed or mislabelled file is still recognised.
What is checked
- The container rules a reader uses to recognise an EPUB at all — the stored mimetype entry has to be first and uncompressed.
- Required metadata: title, identifier and language.
- Manifest entries pointing at files the archive does not contain.
- Reading-order entries pointing at things not in the manifest.
- Documents that a strict XML parser would refuse. EPUB requires draconian parsing, so one unescaped ampersand loses the whole book on a conforming reader.
- Contents entries that lead nowhere, and images with no alternative text.
What it is not
This is not EPUBCheck. EPUBCheck validates the whole specification against its schemas; this checks the shorter list of things that actually break a book in a reader, and explains each one. If you are submitting to a store, run EPUBCheck as well.
Most faults found here are repaired simply by converting the book: reading it and writing it back out rebuilds the package, the navigation and the markup.