EpubMeta

How it works

An EPUB is a ZIP archive with rules. Editing metadata means opening the archive, finding the one XML file that holds the book’s identity, changing it, and packing everything back together correctly. All of that happens in your browser tab.

The pipeline

  1. The file is read locally. The bytes of your EPUB are read with the browser’s own file APIs. No request is made, and nothing is sent anywhere at any point.
  2. The archive is opened in memory. The ZIP is unzipped inside the page.META-INF/container.xml names the package document, usually something likeOEBPS/content.opf.
  3. The package document is parsed. The OPF is XML. Title, author, language and identifiers are Dublin Core elements; series information lives inmeta elements. Both the EPUB 3 form (belongs-to-collection with group-position) and the calibre form (calibre:series) are read, whichever is present.
  4. You edit, and only what you changed changes. Untouched fields keep their original elements and attributes. Series information is written in both encodings at once, because older tools and newer readers each look for a different one.
  5. The book is re-zipped to the letter of the format.mimetype is the first entry and is stored uncompressed; everything else is deflated. That ordering is what makes the file recognisable as an EPUB by reading systems and checkers.
  6. The new file is re-opened and checked before you get it. The page reads the saved bytes back, parses the package document again, and asserts that every edit persisted. The download is blocked if any assertion fails.

What the checks cover

  • The package document parses as XML.
  • Every spine reference matches a manifest item.
  • Every manifest href resolves to a file inside the archive.
  • mimetype is first, stored uncompressed, with the right content.
  • Every edit is re-read from the saved file and compared with what you typed.
  • Every other file in the book is still present, byte for byte.

What the checks do not cover

This is not epubcheck and it is not described as one. Full conformance checking covers hundreds of rules across XHTML, CSS, navigation documents and more. If you need that guarantee, run the saved file through epubcheck as a second step. EpubMeta checks the structural facts it is responsible for, and says so plainly.

Why the book never leaves the device

There is no backend. The site is a set of static files; the editor is a script that runs in the page. If you load the page once, you can disconnect the network and everything — reading, editing, checking, downloading — still works. That is not a policy promise, it is how the tool is built.

DRM-protected books

If a book’s content is encrypted (usually listed in META-INF/encryption.xml), EpubMeta refuses to open it and explains why. It does not attempt to remove DRM, and it never will. Books that only obfuscate embedded fonts are not DRM — the text is not encrypted — and those open normally.