EPUB Metadata: How to Set Title, Author, Language, and Identifiers Correctly

EPUB metadata is the information embedded inside your ebook file that identifies and describes it — the title, author name, language, ISBN, and other structured data that exists separately from the content readers see on screen. When you upload an ebook to Amazon KDP, Apple Books, Kobo, or any other retailer, the platform reads this metadata as part of ingesting your file. It powers catalog listings, search indexing, and screen reader behavior.

Most indie authors pay careful attention to their manuscript content and cover design, then fill in metadata fields quickly as an afterthought. That's a mistake. Wrong or incomplete metadata can break catalog listings, cause text-to-speech to mispronounce your book's language, and prevent proper deduplication between editions. It can also affect how your book is sorted in author catalogs and whether your series books are recognized as belonging together.

This guide covers every major EPUB metadata field, the correct format for each, and the consequences of getting them wrong.

What EPUB metadata is and where it lives

Inside every EPUB file is a package document — typically a file with an .opf extension — that contains the book's metadata in XML format. This file is part of the EPUB container but not part of the reading content. Readers never see it directly; it's read by the ebook store, the reading application, and by assistive technology.

The metadata in the OPF file follows the Dublin Core standard for bibliographic metadata (the dc: prefix on many fields) and the EPUB specification for ebook-specific elements. EPUB 3 introduced additional metadata capabilities beyond what EPUB 2 supported, though the core Dublin Core fields exist in both versions.

When you use a formatting tool like LiberScript, the tool writes this file for you based on the metadata fields you fill in. Understanding what goes into each field helps you fill them in correctly and recognize when something may have been set incorrectly in an existing file.

Core EPUB metadata fields

The table below covers every major metadata field, what it stores, how important it is, the correct format, and a concrete example.

Field	What it stores	Importance	Correct format	Example
dc:title	The book's main title	Critical	Exact title, no subtitle	The Quiet Hours
dc:creator	Author name	Critical	First Last (with file-as: Last, First)	Sarah Monteiro
dc:language	Book's language	Critical	BCP 47 language tag	en
dc:identifier	Unique ID (ISBN or UUID)	Critical	urn:isbn:978... or UUID	urn:isbn:9781234567890
dc:publisher	Publisher/imprint name	Recommended	Your imprint name	Clearwater Press
dc:date	Publication date	Recommended	ISO 8601 (YYYY-MM-DD)	2026-06-16
dc:description	Book description	Optional	Plain text, no HTML	A literary novel about...
dc:subject	Genre/subject classification	Recommended	BISAC code	FIC019000
meta:subtitle	Subtitle	Recommended if applicable	Plain text subtitle	A Novel

Getting the critical fields right is non-negotiable. The recommended fields matter for catalog quality and discovery. Optional fields won't break anything if missing, but filling them in is generally better practice.

The title field: main title only

A common mistake is including the subtitle in the dc:title element. The title field should contain only the main title — the subtitle belongs in a separate meta element.

Correct:

• dc:title: The Quiet Hours
• meta name="calibre:subtitle" (or EPUB 3 equivalent): A Novel

Incorrect:

• dc:title: The Quiet Hours: A Novel

Why does this matter? When retailers sort by title, they sort alphabetically by the title element's content. "The Quiet Hours: A Novel" sorts differently than "The Quiet Hours" and may appear in unexpected places in sorted lists. Some catalog systems also display the title field verbatim — putting the subtitle there creates double-display when the retailer separately shows the subtitle it has on file for your book.

Formatting tools that let you enter title and subtitle as separate fields handle this correctly in the EPUB output. If you're editing metadata manually, keep them separate.

Author name: creator element and file-as attribute

The dc:creator element stores the author's name. For a standard author credit, the role attribute should be set to aut (short for author, following the MARC relator code standard). For books with additional contributors — an editor, illustrator, translator — each has their own dc:creator element with the appropriate role code.

The file-as attribute on the creator element stores the name in Last, First format and is used by catalog systems for alphabetical sorting. Both versions are needed.

<dc:creator opf:role="aut" opf:file-as="Monteiro, Sarah">Sarah Monteiro</dc:creator>

For pen names: use the pen name everywhere in the metadata, not your legal name. The dc:creator is public-facing metadata that appears in catalog listings.

For books with a single author and a foreword by someone else: the foreword writer typically gets a separate creator entry with role wpr (writer of preface) rather than aut. Most indie authors don't include foreword writers in creator metadata at all, which is fine.

Language codes: getting it right matters

The language declaration controls how text-to-speech engines pronounce your book's content. An English book tagged as fr (French) will be read by a French-accented synthesized voice. A bilingual book with sections in two languages needs careful handling.

Use BCP 47 language tags — the standard format for language identification on the web and in ebooks.

Language	Correct BCP 47 tag
English (unspecified)	en
English (US)	en-US
English (UK)	en-GB
French	fr
German	de
Spanish (unspecified)	es
Spanish (Latin America)	es-419
Portuguese (Brazil)	pt-BR
Portuguese (Portugal)	pt-PT
Italian	it
Dutch	nl
Japanese	ja
Chinese (Simplified)	zh-Hans
Chinese (Traditional)	zh-Hant

For most English-language books, en is sufficient. If you're targeting a specific regional market and want text-to-speech to use the correct accent, en-US or en-GB is better. The EPUB spec requires exactly one language declaration for the primary language of the book. See ebook accessibility guide for how language metadata interacts with screen reader behavior.

ISBNs in EPUB metadata

The dc:identifier field stores your book's unique identifier. If you have an ISBN for your ebook, it goes here in the URN format:

urn:isbn:9781234567890

The full 13-digit ISBN with no spaces or hyphens, prefixed by urn:isbn:. This format is required — just the bare ISBN number without the URN prefix will not be recognized correctly by all catalog systems.

ISBNs matter for catalog deduplication. When multiple stores and aggregators list your book, they use the ISBN to recognize that the Kobo listing and the Apple Books listing and the Barnes & Noble listing are all the same title. Without an ISBN — or with a malformed one — you can end up with duplicate listings, merge failures, or catalog inconsistencies.

If you don't have an ISBN: Use a UUID (universally unique identifier) as your identifier. UUIDs are free to generate and ensure your book has a unique identifier even without a commercial ISBN. Many formatting tools generate a UUID automatically. The format is a standard UUID: urn:uuid:550e8400-e29b-41d4-a716-446655440000. A UUID doesn't carry the catalog deduplication benefits of an ISBN, but it satisfies the EPUB spec's requirement for a unique identifier.

Ebook ISBNs are separate from print ISBNs. If you have both a print and ebook edition, each needs its own ISBN. See the ISBN guide for self-publishing for how to obtain and manage ISBNs.

Series metadata in EPUB

Series information — series name and book number — is not standardized in EPUB 2. There's no universal Dublin Core field for it. EPUB 3 introduced the belongs-to-collection meta element, but retailer support for reading this from the EPUB file is inconsistent. Most retailers rely on you entering series metadata through their publishing dashboard rather than pulling it from the EPUB file.

In practice: enter series information when you set up your book on each retailer's platform. Don't rely solely on EPUB metadata to populate series information in store listings. The belongs-to-collection element in EPUB 3 is worth including — some tools and aggregators do read it — but treat it as supplementary to manual entry on each platform. See KDP series setup for how Amazon handles series metadata specifically.

BISAC subject codes

BISAC (Book Industry Standards and Communications) codes are a standardized classification system for book subjects and genres. The dc:subject element in EPUB metadata is the place to include them.

BISAC codes are maintained by the Book Industry Study Group (BISG) and published at bisg.org. Examples:

Genre	BISAC Code
Literary fiction	FIC019000
Mystery	FIC022000
Fantasy	FIC009020
Science fiction	FIC028000
Self-help	SEL000000
Cooking (general)	CKB000000
Biography	BIO000000

Include one to three BISAC codes in your EPUB metadata — the primary genre and any relevant subcategories. Retailers use subject codes for catalog organization and, in some systems, for recommendation algorithms. Most retailers also let you select categories separately in their publishing dashboard, which is where categories that affect browsing placement are usually set.

See KDP categories and keywords for how category selection on Amazon works alongside BISAC codes.

The cover image in OPF metadata

The cover image requires a specific declaration in the EPUB's OPF manifest so that reading systems know which image is the book cover. In EPUB 2, this is done through a <meta name="cover" content="cover-image-id"/> element that references the cover image's manifest ID. In EPUB 3, the cover image item in the manifest should include properties="cover-image" in its attributes.

This declaration is how reading apps identify which image to display as the thumbnail in your library. Without it, some apps may show the first image they encounter (which may be a photo inside chapter one), or display no thumbnail at all.

Formatting tools handle this automatically when you designate your cover image during the export process. If you're troubleshooting a missing or wrong cover thumbnail in an EPUB, the cover image declaration in the OPF is the first place to check.

How formatting tools handle metadata

LiberScript and similar tools expose a metadata form where you enter your title, author, language, ISBN, and other fields before exporting. The tool writes the OPF file based on your entries. This means your formatting tool is only as accurate as the information you provide — if you enter the wrong language code, the EPUB will have the wrong language declaration.

Key points about metadata in formatting workflows:

• Title and subtitle should be entered in separate fields if your tool supports it. If there's only one field, enter the main title only and add the subtitle in a secondary field or after upload.
• Language should be entered as the BCP 47 tag, not as a plain English word ("English" will not work; "en" will).
• ISBN should be entered as the full 13-digit number; the tool will format it as a URN automatically in most cases.
• Author name should match exactly how you want it to appear in store catalogs — the tool should handle the file-as formatting behind the scenes.

After exporting, you can verify your metadata using an EPUB editor like Sigil or a metadata viewer like Calibre, which displays all EPUB metadata fields in a readable format.

Common metadata mistakes and consequences

Mistake	Consequence
Subtitle in the dc:title field	Incorrect sort order; potential double display in store listings
Wrong language code	Text-to-speech mispronounces the entire book
Missing or malformed ISBN	Catalog deduplication failures; duplicate listings across retailers
Author name in Last, First format in the display element	Name appears as "Monteiro, Sarah" in store listings instead of "Sarah Monteiro"
No dc:identifier at all	Violates EPUB spec; some validators and aggregators will reject the file
Missing cover image declaration in OPF	Cover thumbnail missing or wrong in reading apps
Incorrect BISAC code	Miscategorized in catalog systems; potential mismatch with dashboard category selection

Most of these mistakes have no visible symptoms until you see the store listing — which may be days or weeks after upload. Checking your metadata before upload, either through your formatting tool's preview or an EPUB validator, is faster than troubleshooting after the fact.

Frequently asked questions

Does EPUB metadata affect Amazon search results?

Amazon's search algorithm primarily uses the metadata you enter in KDP's publishing dashboard, not what's inside the EPUB file. That said, the title and author you enter in the dashboard should match your EPUB metadata exactly — inconsistencies create catalog confusion. The keywords and categories that affect Amazon search are set in the dashboard, not in the EPUB. See KDP categories and keywords for how to optimize those fields.

Can I update EPUB metadata after uploading my ebook?

Yes. On KDP, Apple Books, and most other platforms, you can update your ebook file by uploading a corrected version. Updates typically take 24–72 hours to propagate to the live listing. The process is: correct the metadata in your source document or formatting tool, export a new EPUB, and upload the new file to replace the existing one. Check each platform's guidelines for how to update a live title.

What's a UUID and do I need one if I have an ISBN?

A UUID (universally unique identifier) is a 128-bit identifier in a standardized format used when no ISBN is available. If you have an ISBN, use that as your identifier — it's more useful for catalog purposes. If you don't have an ISBN and don't plan to get one, include a UUID so your EPUB has a valid unique identifier as required by the EPUB specification. Most formatting tools generate a UUID automatically when no ISBN is provided.

Can my EPUB have two language declarations for a bilingual book?

EPUB 3 allows multiple language declarations when a book contains content in more than one language. The primary language should be declared first, and each additional language can be declared with a secondary dc:language element. For bilingual books, text-to-speech behavior depends on the reading system — not all apps switch pronunciation mid-text based on embedded language spans. For sections of text in a different language from the primary, you can add xml:lang attributes at the paragraph level for more precise language tagging.

Does my EPUB description field show up in store listings?

It depends on the platform. Some aggregators and library catalog systems use the dc:description field from the EPUB. Most major retailers (Amazon, Apple Books, Kobo) use the description you enter in their publishing dashboards and may not pull it from the EPUB file. Including an accurate description in the EPUB metadata is still good practice — it ensures the information is present wherever the file is cataloged or distributed.

The bottom line

EPUB metadata is unglamorous work, but it's the infrastructure that determines whether your ebook is correctly identified, properly sorted, accurately pronounced by screen readers, and reliably deduped across catalog systems. The effort to get it right is small — fill in the fields correctly when you export — and the consequences of getting it wrong compound over the life of your book.

The most important fields are title (main title only), creator (First Last, with file-as Last, First), language (BCP 47 tag), and identifier (URN-formatted ISBN or UUID). Get those four right and you've handled the bulk of what matters. The rest — publisher, date, subject codes, description — improves catalog quality and is worth the few extra minutes to complete.

LiberScript surfaces all key metadata fields in a simple form during export, writes the OPF correctly, and validates the output before you download your file. Get started to format your next ebook, or see pricing to choose your plan.