Does Scrivener produce a valid EPUB directly?

Usually close but not always. Scrivener's compile outputs a structurally valid EPUB for most projects, but it can generate duplicate IDs, redundant CSS, and loose manifest entries that cause failures in strict readers like Apple Books. Always validate before distributing.

What compile format should I use in Scrivener for EPUB 3?

In File → Compile, select the "EPUB 3" format from the Format dropdown. If you need Kindle compatibility, use the "Kindle" format (which produces a MOBI/KF8) or compile to EPUB 3 and send via Send to Kindle.

Why does my chapter title appear twice after compiling?

Scrivener's title prefix settings and your document's first line are both outputting the chapter name. In the compile settings, look at the "Title Options" for your Section Layout and turn off the title prefix, or remove the manual title from the top of each chapter document.

The metadata in my Scrivener EPUB is wrong or missing — how do I fix it?

Open the compiled EPUB in the Epublys metadata editor. You can update title, author, description, and cover without re-compiling from Scrivener. Faster than going back through the compile dialog every time.

Scrivener to EPUB: Compile and Fix

Validate EPUB Files →

Scrivener ships with an EPUB compile option built in, and for basic use it works — chapter breaks, font choices, and a reasonable default layout. What it does poorly is specific enough that most Scrivener-written books end up needing cleanup before they can be uploaded anywhere. Here's what to know going in.

Compile settings that actually matter

Open the compile pane (Cmd+Shift+E on Mac, F7 on Windows) and pick EPUB 3 — not ePub 2. Scrivener's ePub 2 exporter is older and produces less clean output. Focus on three panes:

Format: Start with the default "Ebook" format. Duplicate it before editing anything — Scrivener's built-in formats are read-only and you'll want your own version per project.
Section Types: The part most Scrivener users never touch. Each folder and document in your binder has a section type. Section types map to layouts. Chapters should usually be "Chapter" or "Section"; front/back matter should be "Front Matter" or "Back Matter". Mismatched section types are why your prologue gets formatted like a chapter, or skipped entirely.
Project Metadata: Set this before compile (Project → Metadata). Author, title, description, and language populate the EPUB's metadata. Leaving them blank produces an EPUB with "My Project" as the title and no author.

What Scrivener compiles well

Basic chapter structure with consistent section types
Italic and bold (as long as you used the style toolbar, not a specific font face)
Scene break dividers configured in the Format pane
Automatic table of contents from the binder hierarchy
Straight-through reading order for linear narratives

What Scrivener compiles badly

Epigraphs and block quotes

Scrivener's block quote styling often bleeds into surrounding paragraphs in the EPUB output. Fix by applying a specific paragraph style to the epigraph in Scrivener and overriding it in the compile format's Styles pane. Or: run the compiled EPUB through a CSS-aware cleanup step afterward.

Footnotes

Scrivener's inline footnotes export as HTML footnotes, but the styling is weak and most e-readers render them as awkward inline text rather than pop-ups. For footnote-heavy non-fiction, consider compiling to DOCX and using Pandoc instead (pandoc book.docx -o book.epub handles footnotes cleanly).

Cover image

Scrivener embeds the cover image from your compile settings, but the cover is often placed at a weird size or with odd margins. Compile the EPUB, then swap the cover using the cover replacement tool.

Font embedding

Scrivener can embed fonts into the EPUB, but it embeds the entire font file even if you only use two glyphs, which bloats the file. For body text, set the compile to use default fonts and let readers pick their own. Only embed display fonts if they're critical to the book's identity.

KDP and Apple validation

Scrivener's compiled EPUBs regularly fail EPUBCheck on things like unclosed tags, empty <div> elements, and missing lang attributes. KDP usually accepts them anyway, but Apple Books is stricter. Always run the validator with auto-fix before uploading to Apple.

The standard post-compile cleanup

Validate and auto-fix. Upload the compiled EPUB to the EPUB validator, enable deep scan and auto-fix, download the clean version.
Check metadata. Open the metadata editor. Scrivener sometimes leaves the language code blank (causing accessibility complaints) or uses "Unknown" as the author. Fix these.
Swap the cover. If Scrivener's cover placement is off, use add cover to EPUB with the proper image.
Compress. Run through the compressor to strip Scrivener's overhead and any embedded fonts you don't actually need.

Alternative: Scrivener for writing, something else for export

Some authors use Scrivener purely as a drafting tool and compile to DOCX instead of EPUB. From there, run Pandoc or Calibre for the conversion. This gives you a clean DOCX as an intermediate format you can proof in Word, and the EPUB output is usually cleaner than Scrivener's native compile. Extra step, better result. See the DOCX to EPUB guide for the pipeline.

Validate EPUB — catch Scrivener's validation issues
Edit EPUB metadata — fix author, language
Add or swap EPUB cover
DOCX to EPUB pipeline
Full self-publishing guide

Try it now — free

Check your EPUB for errors, compatibility issues, and compliance. Free EPUB validator supporting EPUB 2 and EPUB 3 standards.