You finished your manuscript in Word. Now you need an EPUB for ebook distribution. The obvious approach — export or convert your DOCX directly to EPUB — seems like it should work. The file has your text, your chapters, your formatting. How hard can it be?
Very. Direct Word-to-EPUB conversion is one of the most common sources of broken, ugly, and rejected ebook files in self-publishing. Understanding why it fails will save you hours of troubleshooting and produce a dramatically better ebook.
Why direct conversion usually fails
Word documents carry an enormous amount of hidden formatting information. Open any DOCX file as XML (which is what it is — a ZIP archive of XML files) and you’ll find thousands of lines of style definitions, revision history, compatibility settings, and formatting overrides that you never intentionally created.
When a conversion tool translates this into EPUB — which is essentially a small website made of HTML and CSS — all of that hidden baggage comes along. The result is an EPUB stuffed with inline styles, redundant class definitions, and formatting artifacts that break rendering on ebook readers.
The problem isn’t the conversion tools. The problem is that Word and EPUB represent documents in fundamentally different ways. Word is page-oriented. EPUB is flow-oriented. Translating between them is not a simple format swap — it’s a structural transformation. This is why tools that import your DOCX and then generate a native EPUB — rather than converting one format to the other — produce consistently better results. Cambric takes this approach: it imports your Word document into a structured manuscript, then generates a clean EPUB from scratch rather than translating Word’s formatting layer.
The seven most common conversion problems
1. Broken paragraph styles
Word lets you format text by selecting it and clicking Bold, changing the font size, or adjusting the indent. These “direct formatting” overrides sit on top of the paragraph style and are almost invisible in Word. But in a converted EPUB, every override becomes an inline style — and inline styles override the ebook reader’s CSS, breaking the reader’s font and size preferences.
The symptom: your EPUB ignores the reader’s font settings. Some paragraphs appear in the wrong size. The text looks inconsistent.
2. Missing or broken chapter breaks
In Word, you might indicate chapters with page breaks, section breaks, or just a heading on a new page. EPUB needs each chapter to be a separate HTML file within the archive. Conversion tools try to detect chapter boundaries automatically, but they often guess wrong — splitting mid-chapter, combining multiple chapters, or creating chapters with no content.
The symptom: your table of contents points to the wrong locations, chapters run together, or you have phantom empty chapters.
3. Phantom formatting
Paste text from a web page or another document into Word, and the source formatting comes along invisibly. Delete a sentence and retype it, and the deleted text’s formatting may persist as a ghost style. These artifacts are invisible in Word but become visible in EPUB as random font changes, unexpected spacing, or text that won’t match the surrounding paragraphs.
The symptom: one paragraph inexplicably uses a different font or size. A line has extra spacing. A section indents differently from the rest.
4. Image problems
Word embeds images in its own internal format and positioning system. Converting to EPUB means extracting those images and re-embedding them in HTML. Common failures: images appear at the wrong size, lose resolution, break out of the text flow, or disappear entirely. Images positioned “in front of text” or “behind text” in Word have no EPUB equivalent.
The symptom: missing images, oversized images that extend off-screen, or images that appear in the wrong chapter.
5. Font embedding issues
Word uses whatever fonts are installed on your computer. Your EPUB can embed fonts, but the conversion tool may not do this automatically — or may embed fonts you don’t have the license to distribute. Many fonts that come with Windows or macOS have licenses that prohibit embedding in ebooks.
The symptom: your EPUB renders in a fallback font on the reader’s device, or the font looks different on Kindle vs. Apple Books vs. Kobo.
6. Table of contents problems
Word’s auto-generated table of contents is a field code that resolves to page numbers. EPUB doesn’t have page numbers — it has anchor links. Conversion tools must rebuild the TOC entirely, and they rely on your heading structure being consistent and correct. If you used manual formatting instead of Heading styles, the converter has nothing to work with.
The symptom: no table of contents in the EPUB, a TOC with wrong or missing entries, or a TOC that lists page numbers (meaningless in an ebook).
7. Special character and encoding issues
Em dashes, curly quotes, ellipses, accented characters — these can break during conversion if the encoding isn’t handled correctly. Older DOCX files or files that have been round-tripped through multiple applications are especially vulnerable.
The symptom: question marks, empty boxes, or garbled characters where special punctuation should be.
The “clean DOCX” approach
If you must convert from Word, the single most effective thing you can do is clean your DOCX before conversion. A clean document converts dramatically better than a messy one.
Strip all direct formatting. Select all text (Ctrl+A), then apply your base paragraph style (usually “Normal” or “Body Text”). This removes every direct formatting override. Yes, you’ll lose your bold and italic — you’ll add those back with character styles. The point is to eliminate the hundreds of invisible overrides accumulated during writing and editing.
Use proper heading styles. Every chapter title should use Heading 1. Every sub-section should use Heading 2. Don’t fake headings by making text bold and large — use the actual built-in heading styles. This gives the converter the structural information it needs to build chapters and a table of contents.
Use paragraph styles, not manual formatting. First-line indents should come from the paragraph style, not the Tab key or the ruler. Block quotes should use a “Block Quote” style, not increased left indent. If you format with styles, the converter can translate those styles into CSS classes. If you format manually, every paragraph gets its own inline style.
Use page breaks for chapter starts. Insert > Page Break (Ctrl+Enter) before each chapter heading. Don’t use multiple Enter keystrokes to push text to the next page — those become empty paragraphs in your EPUB.
Remove headers and footers. EPUB doesn’t have headers or footers. Remove them before conversion so they don’t end up as stray text in your ebook.
Check for hidden content. Turn on paragraph marks (the pilcrow button) and look for stray formatting marks, empty paragraphs, manual line breaks where paragraph breaks should be, and tab characters used for indentation.
Conversion tools compared
Calibre (free, open source)
Calibre is the Swiss Army knife of ebook conversion. It converts between virtually every ebook format, and it’s completely free. The learning curve is steep, and the output requires manual cleanup for professional results. The interface is functional rather than beautiful, and getting clean output from a messy DOCX requires knowing which conversion settings to adjust. Best for technically comfortable authors who want maximum control.
Kindle Create (free, Amazon)
Amazon’s own tool converts DOCX to Kindle format. It’s purpose-built for KDP and produces files that Amazon accepts reliably. The limitation: it only produces Kindle files, not standard EPUB. If you’re exclusive to Amazon, Kindle Create works. If you distribute anywhere else — Apple Books, Kobo, Barnes & Noble — you still need an EPUB. See our Kindle Create comparison for a detailed review.
Reedsy Book Editor (free, web-based)
Reedsy lets you import a DOCX and export an EPUB. The import process strips out most problematic formatting, which is both the strength and the weakness — you get clean output, but you lose any complex formatting you intentionally applied. The editor itself is basic, and the export options are limited. See our Reedsy comparison.
Sigil (free, open source)
Sigil is a dedicated EPUB editor. It doesn’t convert DOCX — you’d convert elsewhere first, then open the EPUB in Sigil to fix problems. It gives you direct access to the HTML and CSS inside the EPUB, which is powerful if you know web technologies and overwhelming if you don’t.
Dedicated formatting tools
Tools like Vellum, Atticus, and Cambric take a different approach entirely. Instead of converting your DOCX into an EPUB, they import your text and let you format it within the tool, then export a properly structured EPUB from scratch. The EPUB is generated natively — it was never a Word document being translated. This sidesteps most conversion problems entirely. See our formatting software comparison for a detailed breakdown.
Testing your EPUB
Never publish an EPUB without testing it. What looks correct in your conversion tool may break on actual reading devices.
EPUB Check. The official validation tool for EPUB files. It catches structural errors, missing metadata, invalid markup, and accessibility issues. Run every EPUB through EPUB Check before uploading. Free at epubcheck.org or as a Java command-line tool.
Kindle Previewer. Amazon’s desktop app simulates how your ebook will look on every Kindle device and app. Even if you upload an EPUB to KDP (which converts it to Kindle format internally), test with Kindle Previewer first. It catches rendering issues that EPUB Check won’t flag.
Side-load to a real device. Send your EPUB to an actual Kindle, Kobo, or iPad and read through it. No simulator perfectly replicates the experience of reading on a real screen. Check the table of contents, verify images, test chapter navigation, and read at least a few chapters at different font sizes.
Test across multiple platforms. An EPUB that looks perfect on Apple Books might break on Kobo, or vice versa. If you distribute wide, test on at least two different reading platforms. Cambric’s EPUB output passes EPUB Check validation out of the box, which eliminates the most common cross-platform rendering issues before you even begin testing.
For specific guidance on ebook formatting standards and best practices, see our ebook formatting guide. For a deeper look at the differences between the two output formats, see our EPUB vs PDF comparison.
Import-then-format beats convert-and-pray
The conversion approach — take a DOCX, run it through a tool, hope the output is acceptable — works for simple manuscripts with clean formatting. For anything more complex, or for authors who want professional-quality output consistently, the better approach is to separate the content from the formatting.
Import your text into a tool designed for book production. Format it there, where the tool understands the constraints of both print and ebook output. Export a clean EPUB that was built as an EPUB from the start, not translated from a page-layout format.
Cambric imports your DOCX, strips the formatting baggage, and lets you apply professional templates designed for ebook output. The EPUB it generates is built natively — proper chapter structure, clean CSS, correct metadata — not converted from a print layout. Import once, format for both print and ebook, and export files that pass validation on every platform. It’s a $109 one-time purchase that runs locally on your desktop, with 20+ templates and Typst-based typesetting that handles both your print PDF and your EPUB from a single project.