Ebook Metadata and Cover Images: Getting Them Right

Metadata is the difference between a book that shows up as "The Great Gatsby by F. Scott Fitzgerald" with a crisp cover image, sorted into the correct series and genre, and one that appears as "Unknown" by "Unknown" with no cover, floating unsorted in a library. Every reading app, every bookstore, every library catalog depends on metadata to organize and display books.

Most authors and publishers treat metadata as an afterthought — fill in the title and author, slap on a cover, done. But incomplete or incorrect metadata has real consequences: books that don't appear in searches, broken series sorting, missing covers on certain devices, and library management that requires manual fixes for every book. Getting metadata right once saves hours of fixing it later.

Dublin Core: The Metadata Foundation

Dublin Core is the international metadata standard (ISO 15836) that EPUB uses for book description. Originally designed for web resource cataloging in 1995, it provides 15 core elements that map well to book metadata:

Three elements are required by the EPUB spec: dc:title, dc:language, and dc:identifier. In practice, you should always include dc:creator (author) and aim to fill in publisher, date, subject, and description. Every missing field is a missed opportunity for discoverability and organization.

EPUB OPF Metadata In Practice

In EPUB, metadata lives in the OPF (Open Packaging Format) file — typically content.opf. Here's a comprehensive real-world example:

<metadata xmlns:dc="http://purl.org/dc/elements/1.1/">
  <dc:title id="title">Dune</dc:title>
  <meta refines="#title" property="title-type">main</meta>
  
  <dc:creator id="author">Frank Herbert</dc:creator>
  <meta refines="#author" property="role" scheme="marc:relators">aut</meta>
  <meta refines="#author" property="file-as">Herbert, Frank</meta>
  
  <dc:language>en</dc:language>
  <dc:identifier id="isbn">urn:isbn:9780441013593</dc:identifier>
  <dc:publisher>Ace Books</dc:publisher>
  <dc:date>1965-08-01</dc:date>
  <dc:subject>Science Fiction</dc:subject>
  <dc:subject>Space Opera</dc:subject>
  <dc:description>Set on the desert planet Arrakis...</dc:description>
  <dc:rights>Copyright 1965 Frank Herbert</dc:rights>
  
  <meta property="belongs-to-collection" id="series">Dune</meta>
  <meta refines="#series" property="collection-type">series</meta>
  <meta refines="#series" property="group-position">1</meta>
  
  <meta property="dcterms:modified">2026-01-15T12:00:00Z</meta>
</metadata>

Creator Roles with MARC Relators

EPUB 3 uses MARC relator codes to distinguish different types of creators. The most common codes for ebooks:

• aut — Author
• edt — Editor
• ill — Illustrator
• trl — Translator
• nrt — Narrator (for audiobook-ebook hybrids)
• aui — Author of introduction
• bkd — Book designer

The file-as property provides the sort name ("Herbert, Frank" instead of "Frank Herbert"). This controls alphabetical sorting in library apps. Without it, books sort by first name, which is wrong for most Western naming conventions. Libraries with thousands of books become unmanageable without correct file-as values.

Series Metadata

EPUB 3 handles series via the belongs-to-collection meta property with a group-position refinement for the book's position in the series. This is how library apps know "Dune" is book 1 of the Dune series.

Calibre uses its own series metadata scheme (stored in <meta name="calibre:series"> and <meta name="calibre:series_index">). Amazon uses EPUB 3's standard series metadata for Kindle display. Apple Books reads both. For maximum compatibility, include both the EPUB 3 standard and Calibre's custom meta elements.

Series metadata directly affects how your book appears in stores. Amazon groups series titles and shows "Book X of Y" on the product page. Kobo and Apple Books use series information for recommendation and navigation. Missing series metadata means your sequel doesn't link to the original.

MOBI/AZW Metadata: EXTH Records

MOBI and AZW files store metadata in EXTH (Extended Header) records — a flat key-value system using numeric type codes. The most relevant records:

When converting MOBI to EPUB (convert here), Calibre maps these EXTH records to Dublin Core elements automatically. The reverse conversion (EPUB to MOBI) does the same mapping in the opposite direction. The main metadata loss in conversion is EPUB 3's extended refinements (role types, file-as, series position) which have no direct MOBI EXTH equivalent.

Cover Image Specifications by Platform

Cover requirements vary by platform, but designing for the largest (Amazon KDP) covers all of them:

Platform-Specific Requirements

The universal safe spec: 2560x1600 pixels, RGB, JPEG, quality 95. This meets or exceeds every platform's requirements. Some platforms (Apple, B&N) prefer taller aspect ratios (4:3 or 5:7), but a 1.6:1 cover works everywhere — platforms that expect taller covers add letterboxing or crop slightly.

Embedding the Cover in EPUB

The cover image in EPUB is referenced in the OPF manifest with the cover-image property:

<item id="cover-img" href="images/cover.jpg" 
      media-type="image/jpeg" properties="cover-image"/>

Additionally, most readers expect a cover page — an XHTML document that displays the cover image, included as the first spine item:

<!-- cover.xhtml -->
<html xmlns="http://www.w3.org/1999/xhtml">
<head><title>Cover</title>
<style>body{margin:0;padding:0;text-align:center}
img{max-width:100%;max-height:100%}</style>
</head>
<body>
  <img src="images/cover.jpg" alt="Book Cover"/>
</body>
</html>

Include both the cover-image manifest property AND a cover page in the spine. Some readers use one, some use the other. Including both guarantees the cover displays everywhere. The cover image file should be the same high-resolution image you upload to the store — don't embed a low-res thumbnail in the EPUB and a high-res version on the store page.

Subject and Genre Classification

Genre metadata affects where your book appears in store categories and search results. Different platforms use different classification systems:

• Amazon KDP: Uses BISAC (Book Industry Standards and Communications) codes and their own category tree. You select up to 3 categories during KDP publishing. Amazon also reads dc:subject from the EPUB for keyword matching
• Apple Books: Uses their own category taxonomy (similar to BISAC). Categories are selected during submission
• EPUB metadata: dc:subject accepts freeform text. Include genre terms, subgenre terms, and relevant keywords. Multiple dc:subject elements are allowed
• ONIX: The publishing industry standard for metadata exchange. Uses BISAC, BIC, or Thema subject codes. Required by distributors like IngramSpark

For EPUB dc:subject, be specific: "Science Fiction," "Hard Science Fiction," "Space Opera" is better than just "Fiction." Library management software and some stores use these terms for filtering and recommendation. The more precise your subjects, the better your book gets categorized.

Tools for Managing Ebook Metadata

Calibre is the most comprehensive metadata tool. Its metadata editor reads and writes metadata for EPUB, MOBI, AZW3, PDF, and FB2. You can: edit all Dublin Core fields, manage series and series index, set custom columns for personal tracking, download metadata from online sources (Amazon, Google Books, Open Library), bulk-edit metadata across multiple books, and embed/replace cover images. For a library of hundreds or thousands of ebooks, Calibre's metadata management is indispensable.

Sigil provides a metadata editor for EPUB files with direct access to the OPF XML. It's more precise than Calibre for EPUB-specific metadata but doesn't handle other formats.

exiftool can read metadata from EPUB, MOBI, and PDF files from the command line: exiftool book.epub. Useful for quick metadata inspection without opening a GUI application.

Metadata Survival During Format Conversion

When converting between ebook formats, metadata transfer is generally good but not perfect:

• EPUB to MOBI/AZW3: Title, author, publisher, date, description, ISBN, subject, and cover transfer cleanly. Series metadata depends on the converter — Calibre handles it well. MARC relator codes (author roles) are lost
• MOBI to EPUB: EXTH records map to Dublin Core. Cover image transfers if properly indexed. Series metadata in MOBI is non-standard and may require manual verification after conversion
• FB2 to EPUB: FB2's rich metadata (title-info, document-info, publish-info) maps well to Dublin Core. Genre taxonomy converts to dc:subject text. Cover and series metadata transfer
• PDF to EPUB: PDF metadata (Title, Author, Subject, Keywords in the document properties) transfers to Dublin Core. But PDF metadata is often incomplete or absent — many PDFs have no metadata at all

After any format conversion, always verify metadata in the output file. Open in Calibre, check title/author/series/cover, and fix any missing or mangled fields before distributing.

Metadata is infrastructure. It's not glamorous, and readers never see it directly. But it determines whether your book shows up in searches, sorts correctly in libraries, displays its cover on store pages, and links to its sequels. Five minutes of thorough metadata entry during EPUB creation saves hours of troubleshooting later.

The minimum viable metadata is: title, author (with file-as sort name), language, identifier (ISBN or UUID), cover image, and series information if applicable. The maximum useful metadata adds publisher, date, multiple subjects/genres, description, and rights. Fill in everything you can — storage is free, and every field improves discoverability.