EPUB metadata is the book's identity record: title, author, publisher, publication date, language, and a unique identifier. It lives apart from the body text, and reading apps use it to shelve the book, sort it, and return it in search results. If it is empty or wrong, a perfectly typeset body still will not behave like a book.
This article covers what to enter, where those values end up inside the EPUB file, and when a missing entry becomes visible. Building the table of contents and the overall anatomy of an EPUB file are covered in separate articles; this one stays on metadata.
Six practical fields to settle for self-publishing
EPUB 3 keeps the book's identity in the package document's metadata. The specification requires an identifier, title, and language. Author, publisher, and publication date are optional metadata, but keeping all six together is useful for managing a self-published title and its revisions.
- Title — the name on the shelf. If it is part of a series, decide here whether the volume number belongs in the title
- Author — if you publish under a pen name, keep the exact spelling identical across every later volume
- Publisher — your own name or your imprint. It can be blank, but the colophon looks bare without it
- Publication date — the date you release. If you plan revisions, track it together with a version number
- Language — Japanese, English, and so on. Reading apps use it for display and word-breaking decisions
- Unique identifier — a value that points at this book and no other. Even without an ISBN, self-published titles need one that does not collide
The two that most often stay blank are language and the identifier. While you are absorbed in the body text, they slip past you all the way to the end.
Doing it by hand means learning a file format
If you assemble an EPUB yourself, the metadata goes into a separate file with a fixed syntax. Title here, author there — the positions are defined, and so are the ordering and punctuation rules.
Converters are another route, but their metadata fields and generated pages vary. If the tool you choose does not expose a field you need, you have to supply that value in a separate step.
Neither approach is beyond anyone. For a single book, you can look up the syntax and copy it.
The real cost is duplication, not syntax
The syntax is not what wears you down. If you choose to provide both machine-readable metadata and a human-readable publication page, shared values have to be managed in two forms.
The EPUB specification defines machine-readable metadata in the package document. A colophon is not a second required metadata store; it is an optional content page created by the publisher or the production tool. When you include one, it repeats values such as title, author, and publication date in a form readers can see.
By hand, the sequence goes like this.
- Write title, author, and date into the machine-readable record
- Build a separate colophon page and write the same title, author, and date as readable prose
- Shift the publication date by one day, and edit both places
- Forget one of them
Every revision multiplies the chance of that last step. Readers usually notice the colophon, so you end up with a shelf entry that is current and a colophon still showing an old date.
A missing field is the other cost. Depending on the workflow, a missing value may not surface until a late check. A pre-export check keeps metadata corrections separate from manuscript revisions.
In rune Studio the values are entered once and appear in both places
rune Studio holds title, author, publisher, writing direction, output file name, and language as series information, and keeps the publication date and version per volume. Select your manuscript files, export, and both the machine-readable record and the reader-facing colophon are produced from those same values.
I checked this by driving the development build from the command line. With the title set to a sample work, the author, the publisher, Japanese as the language, version 1.0, and 5 September 2026 as the publication date, the exported EPUB carried title, author, date, language, publisher, and a unique identifier in its record, while the colophon page listed, automatically:
- the title as a heading
- the publication date with the version number
- the author line
- the publisher line
- a copyright notice
- a note that the work is fiction
Nothing was typed twice. Changing the publication date from 5 September to 1 October and exporting again put the new date in both the machine-readable record and the colophon.
One thing I noticed: raising the version from 1.0 to 1.1 and exporting left the previous version's line in the colophon and added the new one beneath it. Successive revisions accumulate a publication history there. Whether that is intended was not something I could confirm, so it is worth exporting once and reading the colophon before you issue a revision.
The language setting changes how the colophon is written. Exporting with the language set to English produced a colophon headed "Colophon" and a publication line reading "September 5, 2026 eBook ver 1.0 published" — English date order, English wording. Leave the language on Japanese for an English book and that line keeps its Japanese shape.
Missing fields are named before the export runs
Try to export with the metadata still blank and rune Studio stops and names what is missing. Running that check on an empty setup returned three items — no title, no author name, no output file name — and no EPUB was written. After setting title, author, publisher, language, writing direction, and output file name as the full series information, the same check returned zero findings.
A three-item list before export tells you exactly which settings to revisit. Store acceptance remains a separate check.
Who this suits, and who is fine without it
It suits anyone who re-releases the same book repeatedly. If your rhythm is fix a typo, export, add a scene, export, then a workflow where the metadata cannot drift out of sync is worth having. The same applies to a series where the date and version change per volume.
For a single book released once, doing it by hand can be reasonable. If you already have a conversion routine that fills in and tracks metadata, there is little reason to switch.
One limit is worth stating plainly. What was verified here is that the entered metadata appears in the EPUB package metadata and rune Studio's generated colophon, and that the three tested missing fields are reported before export. General EPUB conformance, store acceptance, and rendering in each reading app were outside this check. Verify those separately against current store guidance and the reading apps you plan to support.
Summary
- EPUB requires an identifier, title, and language; authors often track author, publisher, and publication date alongside them
- Package metadata and an optional reader-facing colophon are different things, even when they share values
- rune Studio uses the entered work information to produce package metadata and its generated colophon, adapting the latter to the chosen language
- The empty setup returned three missing-field findings; the full series setup returned zero before export
For a first step, write your book's six metadata fields on paper. If two or more are blank, filling those in will get you to a finished book faster than another pass over the manuscript. See the current product scope on the Rune Studio product page.
