Step-by-step guide to converting Word documents to EPUB format. Learn about heading hierarchy, image optimization, TOC generation, metadata, cover image requirements, and validation.
Microsoft Word is where most books begin. Whether you are a novelist finishing your manuscript, a business author publishing a professional guide, or a freelancer creating an ebook lead magnet, Word is the starting point for the vast majority of published ebooks worldwide.
The challenge is getting from a Word document to a properly structured EPUB file that looks good on every ebook reader. Word is a word processor, designed for printed pages. EPUB is a web-based format, designed for screens of all sizes. Bridging that gap requires understanding how both formats work and where they diverge.
This guide walks you through the entire process, from preparing your Word document for clean conversion to validating and distributing your finished EPUB.
The single most important factor in EPUB quality is how well your Word document is structured before conversion. A document that relies on manual formatting (bold text for headings, manual page breaks, tab characters for indentation) will produce a poorly formatted ebook. A document that uses Word's built-in styles will convert cleanly.
Headings define the structure of your ebook. They determine chapter breaks, table of contents entries, and navigation landmarks. You must use Word's built-in heading styles:
Never create "fake headings" by making text bold and large. EPUB readers use the heading hierarchy for navigation. Fake headings are invisible to the navigation system and will not appear in the table of contents.
Use the Normal style for body paragraphs. Do not override Normal with manual font or size changes. Instead, modify the Normal style itself if you want different formatting during writing.
For special paragraph types, use or create named styles:
Before conversion, search for and remove these common issues:
| Problem | How to Find | Fix |
|---|---|---|
| Manual line breaks | Find ^l (Shift+Enter) | Replace with ^p (paragraph break) or remove |
| Double paragraph marks | Find ^p^p | Replace with single ^p (or use paragraph spacing) |
| Tab indentation | Find ^t at paragraph start | Remove; use First Line Indent in paragraph style |
| Manual page breaks | Find ^m | Use Section Break or Heading style page break |
| Multiple spaces | Find two or more spaces | Replace with single space |
| Forced font changes | Select All, apply Normal style | Reapply named styles only |
Pro Tip: Use Word's Find and Replace (Ctrl+H) with special characters to automate cleanup. Enable "Use wildcards" for advanced patterns. Cleaning your document before conversion saves hours of manual EPUB editing afterward.
Word headers, footers, and page numbers do not convert to EPUB because ebooks do not have fixed pages. The reader device controls pagination dynamically. Remove all headers and footers before conversion to avoid garbled text appearing at the beginning of chapters.
Images are the most common source of ebook problems. Too large and your ebook exceeds platform upload limits or loads slowly. Too small and images look pixelated. Wrong format and some readers cannot display them.
| Parameter | Recommended Value | Maximum |
|---|---|---|
| Format | JPEG (photos), PNG (diagrams) | No GIF animation |
| Color space | sRGB | Not CMYK |
| Resolution | 72-150 DPI | 300 DPI |
| Max width | 1600 px | 4000 px (platform-dependent) |
| Max height | 2400 px | 4000 px |
| JPEG quality | 75-85% | 95% |
| File size per image | Under 1 MB | 5 MB |
| Total images | Keep under 50 MB | 650 MB (Amazon limit) |
For clean EPUB conversion:
Every ebook needs a cover image. This is both a marketing asset (the thumbnail readers see in store listings) and a technical requirement (platforms reject uploads without covers).
| Platform | Minimum Size | Recommended Size | Aspect Ratio | Format |
|---|---|---|---|---|
| Amazon KDP | 625 x 1000 px | 2560 x 1600 px | 1.6:1 | JPEG/TIFF |
| Apple Books | 1400 x 1873 px | 2560 x 1600 px | 1.6:1 | JPEG/PNG |
| Google Play | 1600 x 2400 px | 2560 x 1600 px | 1.5:1 | JPEG/PNG |
| Kobo | 1400 x 1873 px | 2560 x 1600 px | 1.6:1 | JPEG/PNG |
Do not embed the cover image inside your Word document as a regular image. Most conversion tools have a separate cover image option that places the cover in the correct location within the EPUB package.
With your Word document properly prepared, it is time to convert. There are several reliable approaches.
The fastest method for most users:
.docx fileOur converter preserves heading structure, images, links, and table of contents from properly formatted Word documents.
Calibre is a free, open-source ebook management application with powerful conversion capabilities:
C)Pro Tip: In Calibre's conversion dialog, open the Heuristic Processing tab and enable "Unwrap lines" if your source document has hard line breaks within paragraphs. This is common with PDF-to-Word conversions and will break your EPUB layout if not corrected.
For developers and technical users, Pandoc provides precise control:
pandoc input.docx -o output.epub \
--epub-cover-image=cover.jpg \
--toc \
--toc-depth=2 \
--epub-chapter-level=2 \
--metadata title="My Book Title" \
--metadata author="Author Name" \
--metadata lang="en" \
--css=ebook-styles.css
Key flags:
--epub-cover-image: Sets the cover (does not embed it as a regular page)--toc: Generates a table of contents--epub-chapter-level=2: Splits the EPUB into separate files at Heading 2 level (improving reader performance)--css: Applies custom CSS for typography and layoutSigil is a dedicated EPUB editor that can import Word documents (via a plugin) and gives you direct access to the EPUB's internal HTML, CSS, and metadata:
Sigil is the best option when you need fine-grained control over the EPUB output, particularly for fixing conversion issues that other tools cannot handle.
Ebook metadata determines how your book appears in store listings, library catalogs, and reader apps. Complete metadata improves discoverability and professionalism.
| Field | Description | Example |
|---|---|---|
| Title | Book title | "The Developer's Guide to APIs" |
| Creator/Author | Author name(s) | "Jane Smith" |
| Language | ISO 639-1 code | "en" (English), "fr" (French) |
| Identifier | Unique ID (ISBN preferred) | "978-0-123456-78-9" |
| Publisher | Publishing entity | "Tech Press LLC" |
| Date | Publication date | "2026-02-19" |
| Description | Book summary (for store listings) | 200-500 word description |
| Subject | Categories/genres | "Computer Science", "Web Development" |
| Rights | Copyright statement | "Copyright 2026 Jane Smith" |
| Cover | Reference to cover image | Points to cover.jpg in manifest |
E)Create a YAML metadata file (metadata.yaml):
---
title: "The Developer's Guide to APIs"
author: "Jane Smith"
date: "2026-02-19"
lang: "en"
publisher: "Tech Press LLC"
rights: "Copyright 2026 Jane Smith. All rights reserved."
description: |
A comprehensive guide to designing, building, and consuming
REST and GraphQL APIs. Covers authentication, rate limiting,
versioning, and best practices for production deployments.
---
Then reference it in your conversion command:
pandoc metadata.yaml input.docx -o output.epub \
--epub-cover-image=cover.jpg \
--toc
A well-structured table of contents is essential for ebook navigation. There are two types of TOC in an EPUB:
This is a visible page within the ebook that readers can navigate to. It should list all chapters and major sections with clickable links. Most conversion tools generate this automatically from your heading structure.
This is a machine-readable navigation file used by the reader device to populate its built-in navigation panel (the panel that appears when you press the "Go To" or "Table of Contents" button on a Kindle or e-reader). EPUB 3 uses a nav element in XHTML; older EPUB 2 uses a separate NCX file.
Both TOC types should be generated automatically if your Word document uses proper heading styles. If they are not, your heading structure has problems that need to be fixed before conversion.
Before submitting to any platform, validate your EPUB to catch errors that would cause rejection or rendering problems.
EPUBCheck is the official EPUB validation tool, maintained by the W3C:
# Install (requires Java)
# Download from github.com/w3c/epubcheck/releases
# Run validation
java -jar epubcheck.jar your-book.epub
EPUBCheck reports three severity levels:
| Error | Cause | Fix |
|---|---|---|
Missing dc:title | No title in metadata | Add title metadata |
| Invalid content type | Wrong file extension in manifest | Correct MIME types in OPF |
| Undeclared resource | Image/font referenced but not in manifest | Add to manifest or remove reference |
| Invalid XHTML | Malformed HTML in content | Fix HTML syntax errors |
| Missing alt text | Images without alt attributes | Add descriptive alt text |
| CSS parsing error | Invalid CSS properties | Fix or remove unsupported CSS |
Validation catches technical errors but not visual problems. Always test your EPUB on actual reading devices or their official previewing tools:
Test at multiple font sizes, in both light and dark mode, and on different screen sizes to catch layout issues.
Cause: Character encoding mismatch. Your Word document may use special characters (curly quotes, em dashes, non-Latin characters) that are not properly encoded during conversion.
Fix: Ensure your Word document is saved as .docx (not .doc). The older .doc format uses a different encoding that is more error-prone. If characters are still garbled, open the EPUB in Sigil and set the content encoding to UTF-8.
Cause: Images were not embedded in the Word document (they were linked to external files) or they were inserted as floating objects instead of inline.
Fix: In Word, select each image and set wrapping to In Line with Text. If images are linked, right-click and choose Save as Picture, then reinsert them as embedded images.
Cause: Custom fonts from your Word document were not embedded in the EPUB, so each reader substitutes its default font.
Fix: Embed fonts in the EPUB using Sigil or Calibre's font embedding option. Note that font licensing may restrict embedding. Use freely licensed fonts (Google Fonts, Adobe Fonts with web embedding rights) to avoid legal issues.
Cause: Heading styles were not used, or headings were applied inconsistently.
Fix: Go back to your Word document and apply proper heading styles (Heading 1, Heading 2, etc.) to all chapter titles and section headings. Then reconvert.
Cause: High-resolution images, embedded fonts, or unnecessary media.
Fix: Optimize images before conversion (see Step 2). Subset fonts to include only used characters. Remove any embedded audio, video, or unused images. Our guide on image compression without quality loss covers image optimization techniques.
For authors who want precise control over ebook typography, custom CSS lets you style every element:
/* Base typography */
body {
font-family: "Georgia", "Times New Roman", serif;
line-height: 1.5;
margin: 1em;
}
/* Chapter titles */
h1,
h2 {
font-family: "Helvetica Neue", "Arial", sans-serif;
text-align: center;
margin-top: 3em;
margin-bottom: 1.5em;
page-break-before: always;
}
/* First paragraph of chapter (no indent, drop cap) */
h2 + p {
text-indent: 0;
}
h2 + p::first-letter {
font-size: 3em;
float: left;
line-height: 1;
padding-right: 0.1em;
}
/* Body paragraphs */
p {
text-indent: 1.5em;
margin: 0;
}
/* Block quotes */
blockquote {
margin: 1em 2em;
font-style: italic;
border-left: 3px solid #ccc;
padding-left: 1em;
}
/* Images */
img {
max-width: 100%;
height: auto;
}
Apply this CSS in Calibre's conversion settings (under Look & Feel > Styling), as a --css file in Pandoc, or by editing the EPUB's stylesheet directly in Sigil.
Before uploading to publishing platforms, verify:
For a broader understanding of ebook format choices, see our best ebook format guide. If you need to convert your EPUB to PDF for a different distribution channel, our guide on converting EPUB to PDF covers that process. And for detailed format comparisons, our EPUB vs PDF analysis helps you decide which format to prioritize for your audience.
Priya Patel
UX researcher and technical writer exploring document accessibility, font technology, and cross-platform compatibility.