EPUB metadata

Opens an EPUB and shows what is inside its OPF descriptor: author, contributor, publisher, date, description, tags, the required dc:title / dc:language / dc:identifier — and crucially every identifying marker that stores and library managers silently embed into the book: iTunesMetadata.account (the purchaser’s Apple ID, literally an email address in cleartext), calibre:timestamp and calibre:title_sort (Calibre stamps when each book was added to your personal library), MOBI-ASIN and KFX (Kindle), BN_BookID (Barnes & Noble Nook), adobe- tags (Adobe Digital Editions), Sigil-version (editor signature), kobo markers. Then strips whichever groups you select via checkboxes: identifying fields (creator/contributor/publisher/source), store markers, dates, descriptive fields (description/subject/rights). EPUB-spec-required fields are preserved: at least one dc:title, dc:language and primary dc:identifier — otherwise the book would stop opening in most readers. The primary identifier is optionally replaced with a random urn:uuid:*: this preserves EPUB validity but severs the link between the book and its original ISBN/ASIN. Other extra identifiers (additional dc:identifier entries with ASIN/MOBI-ASIN/Adobe-ISBN/Kobo schemes) are removed. Technically: an EPUB is a ZIP container with a mandatory first entry named mimetype (uncompressed STORED, no extra field — this invariant is critical, without it readers treat the file as corrupt), a META-INF/container.xml pointing to the OPF, and the .opf file itself with metadata in Dublin Core form. The ZIP writer outputs a minimal valid archive respecting the mimetype-first invariant. Several files at once produce a ZIP. Processing happens in the browser: the book never leaves your device by a single byte.