calibre DOCX Conversion Formatting Broken: How to Fix It

When a DOCX file looks correct in Microsoft Word but develops broken headings, excessive spacing, misplaced images, missing footnotes, or inconsistent styles after calibre converts it, the problem usually falls into one of three categories. The source document contains formatting that does not translate cleanly into an e-book, a conversion option is changing the document structure, or the resulting e-book is being displayed differently by a particular viewer or device. The safest solution is to test a copy, simplify the DOCX structure, and change one conversion setting at a time. This guide focuses on DOCX source cleanup and the calibre options directly related to formatting.

1. Confirm the Symptom With a Small Safe Test

Do not begin by reinstalling calibre, deleting your library, or repeatedly converting the entire manuscript with different settings. First establish whether the formatting problem comes from the DOCX file, the conversion configuration, or the program used to display the result.

1.1 Work With a Copy of the Document

Make a copy of the original DOCX before changing anything. Give it a clear name such as book-conversion-test.docx and store it in a normal local folder. Avoid opening and editing the copy directly inside a calibre library folder because calibre manages the files stored there.

Open the copied document in Word or another DOCX-compatible editor. Remove all but a representative sample containing:

• A title and two heading levels
• Several normal body paragraphs
• A bulleted or numbered list
• One image with its caption
• At least one footnote or endnote
• Any section that currently develops incorrect spacing

Save this reduced sample as a new DOCX, add it to calibre as a separate book, and convert it to EPUB. EPUB is a useful diagnostic target because calibre can display and edit its underlying HTML and CSS.

Success looks like this: if the sample converts correctly, the original document probably contains a problematic section, object, or accumulated formatting. Stop changing global calibre settings and identify what is present in the full document but absent from the sample. If the sample is also broken, continue with source cleanup and conversion checks.

1.2 Compare More Than One Viewer

Open the converted EPUB in calibre's E-book viewer, then test it in the reading application or physical device you normally use. A layout that fails only in one reader may be a reader limitation rather than a conversion failure.

Reflowable e-books do not preserve Word pages exactly. Page boundaries, line endings, margins, image placement, and the amount of text visible on a screen change with the reader's font size and display dimensions. A different page count is therefore normal. Missing text, incorrect heading hierarchy, overlapping images, broken note links, and unexplained blank space are not.

Success looks like this: headings, paragraphs, images, and note links remain logically correct in multiple viewers, even if page breaks and line wrapping differ. Once the structure is correct, stop trying to reproduce Word's printed page layout in EPUB.

2. Fix the DOCX Source Before Changing Advanced calibre Settings

If calibre DOCX conversion formatting is broken, source cleanup usually produces a more reliable result than stacking conversion overrides. calibre converts the DOCX into an internal HTML-based representation before creating the requested e-book. Formatting that depends on manual tabs, repeated spaces, text boxes, floating objects, or printer-oriented page positioning may not survive that process as intended.

2.1 Use Word Styles Instead of Manual Formatting

Apply named paragraph styles to structural elements. Use Heading 1 for chapter titles, Heading 2 for major subsections, and Heading 3 only where another level is genuinely needed. Use Normal or a dedicated body style for ordinary paragraphs.

Do not create headings only by enlarging the font, applying bold, changing the color, or centering a Normal paragraph. Those changes describe appearance but may not identify the paragraph as a heading. Similarly, avoid creating paragraph spacing with repeated blank paragraphs or manual line breaks.

For body text, modify the relevant style so it consistently defines alignment, first-line indentation, and space before or after. Remove tabs or groups of spaces used to imitate indentation. In Word, enable formatting marks so you can see paragraph marks, tabs, manual line breaks, and repeated spaces.

Success looks like this: selecting a chapter title shows a Heading style in Word, selecting body text shows the intended body style, and paragraphs no longer depend on blank lines or tabs for layout. Convert again at this point. If spacing and hierarchy are correct, stop changing settings.

2.2 Verify Headings and Table of Contents Detection

calibre can generate a table of contents from Word's built-in heading styles. Before conversion, inspect the document's heading outline. In desktop Word, the Navigation pane provides a convenient test: headings that appear there are usually structurally marked, while visually styled paragraphs that do not appear may require a real heading style.

Keep the hierarchy logical. A chapter should not jump from Heading 1 directly to Heading 4 without a reason. Remove heading styles from decorative text that should not become a table of contents entry, such as running labels, quotations, or caption text.

During conversion, review the Table of Contents and Structure detection sections. Start with calibre's defaults rather than adding custom XPath expressions immediately. If the source headings are correctly styled, the default detection is often enough. If the DOCX contains an outdated or unwanted table of contents, consider forcing an automatically generated table of contents or manually fine-tuning the ToC after conversion.

Success looks like this: the converted book's navigation panel lists chapters and subsections in the correct order, and each entry opens the expected location. Stop modifying detection rules when the ToC and visible heading structure are both correct.

2.3 Normalize Paragraph and Line Spacing

Mixed spacing often comes from pasted text carrying several style definitions or direct paragraph formatting. Select affected paragraphs in the DOCX and reapply the intended body style. Where necessary, clear direct formatting first, then apply the style again.

Inspect the paragraph settings for space before, space after, line spacing, and first-line indentation. Avoid combining a first-line indent with tabs at the beginning of each paragraph. Also remove empty paragraphs inserted between every body paragraph. An e-book stylesheet can create visual separation without adding empty content.

If only a few paragraphs remain wrong after conversion, compare their Word style names and settings with nearby paragraphs that convert correctly. A seemingly identical paragraph may use a duplicated style or contain direct formatting inherited from another document.

Success looks like this: paragraphs use predictable spacing at different viewer font sizes, with no isolated blocks that are unexpectedly indented, centered, or separated by large gaps.

2.4 Prepare Images and Captions for Reflow

Images are most dependable when inserted as normal inline objects rather than floating shapes anchored beside text. Floating placement, text wrapping, grouped shapes, SmartArt, and text boxes are designed primarily for page layout and may move or flatten during conversion.

In the DOCX copy, change important illustrations to an inline layout where practical. Crop and resize images before conversion instead of relying on extreme Word scaling. Use standard formats such as JPEG or PNG, and confirm that the original image has enough resolution for the intended screen size.

Place a caption in its own paragraph immediately below or above the image. Apply a consistent caption style instead of putting the caption in a floating text box. If an image and caption must remain together, test the actual result on the target reader because reflowable formats cannot guarantee that both will always remain on the same screen.

Success looks like this: every image appears once, maintains a sensible aspect ratio, scales within the reading area, and remains next to its caption in normal reading order. If that works across the target readers, do not keep changing output profiles in pursuit of identical pagination.

2.5 Check Footnotes and Endnotes

Create notes with Word's built-in footnote or endnote feature. Superscript numbers typed manually are only text and do not create navigable note relationships. Likewise, notes assembled with bookmarks, text boxes, or manually copied numbers can lose their connections.

Test note references in the converted EPUB. The reference should open or move to the note text, depending on the reader, and the return link should take the reader back when supported. Reader applications can present notes differently, so a popup in one application and a linked note section in another can both be valid results.

If notes disappear or become plain text, copy one failing note and one surrounding paragraph into the small test DOCX. Recreate the note with Word's built-in command and convert again. If the recreated note works, repair the note structure in the source rather than adding conversion styling.

Success looks like this: every note marker leads to the correct note, the note text is readable, and numbering remains consistent. Stop when the links work, even if the visual presentation differs between reading applications.

2.6 Save a Clean DOCX

Documents assembled from many older files can accumulate duplicate styles, compatibility settings, tracked changes, comments, and hidden objects. Accept or reject tracked changes as appropriate, remove comments that should not be published, and delete obsolete text boxes or drawing objects.

Use Save As to create a fresh DOCX rather than repeatedly overwriting an old file. If the document originated as DOC, RTF, an online editor export, or a scanned conversion, open it in a current DOCX-capable word processor, inspect it carefully, and save a new DOCX copy.

If corruption is suspected, create a blank document and copy the content into it in sections. Avoid copying the final paragraph mark if doing so would carry damaged document-level formatting. Reapply the core styles in the new file and save it locally.

Success looks like this: the clean DOCX opens without repair warnings, its heading outline is correct, images display normally, and the reduced test converts consistently.

3. Check calibre Conversion Settings Related to the Symptom

Once the DOCX is structurally clean, inspect only the conversion options that can affect the observed problem. Resetting unrelated metadata, server, email, device-detection, or library settings will not repair malformed headings or source spacing.

3.1 Confirm the Input and Output Formats

Open Convert books and confirm that DOCX is shown as the input format. Select the output format intentionally. EPUB is generally the easiest format for diagnosis and cleanup. Use AZW3 or another format only when it is appropriate for the intended device or application.

Check that you are converting the newest DOCX revision. calibre copies files when they are added, so editing an external DOCX does not necessarily update the copy already stored with the calibre book. Add the revised DOCX to the existing book record as a replacement format, or create a clearly named test record.

Success looks like this: the conversion dialog names the correct DOCX input, and the source content matches the latest saved revision.

3.2 Review Look and Feel Overrides

Conversion rules under Look and feel can deliberately alter fonts, paragraph spacing, indentation, justification, and CSS. Custom style transformations, filtered CSS properties, or extra CSS may override formatting that was correctly imported from the DOCX.

For a clean test, remove custom Extra CSS and disable style transformations you do not recognize. Do not enable broad filtering simply because the document looks untidy. Removing all margins, font sizes, or text alignment can solve one symptom while damaging captions, quotations, headings, and lists.

If the clean conversion works, restore only the customization you actually need, one rule at a time. Reconvert after each change.

Success looks like this: the default conversion preserves the document's logical hierarchy and consistent styling. Once you identify a custom rule that recreates the defect, leave that rule disabled or narrow its scope.

3.3 Check Page Setup and Output Profile

Page setup controls margins and helps calibre interpret measurements for the source and target screen. Choose a specific output profile when you know the destination device. Otherwise, use an appropriate generic profile rather than selecting an unrelated device simply to change image or text size.

An output profile cannot turn a reflowable EPUB into a fixed Word page. It may affect screen-size assumptions, image scaling, and font rescaling. If margins or images look wrong only on one small-screen device, test the matching profile. If the book will be read across many devices, a generic modern e-ink profile is usually a better baseline.

PDF requires separate attention because it is page-based. Confirm the PDF page size and margins in the PDF output settings. A profile intended for a small reader screen can produce very different pages from a standard print-oriented page size.

Success looks like this: images fit within the display, text has usable margins, and font sizing is reasonable on the target device. Stop once the content is readable and structurally correct.

3.4 Rule Out Saved Conversion Settings

calibre can remember conversion options for an individual book. That means one DOCX may continue producing an unusual result even when another document converts normally.

Open the conversion dialog for the affected book and restore conversion settings to defaults where available. Then apply only the target format and necessary page profile. If a new calibre book record converts differently from the existing record, compare the per-book options rather than assuming the library is damaged.

Plugins are not a common cause of DOCX layout problems, but a plugin or custom workflow that modifies metadata, formats, or conversion behavior can complicate testing. Temporarily repeat the conversion without optional processing steps. Do not remove plugins randomly if the same symptom occurs in a plain default conversion.

4. Check Source Location and Reader Limitations

Operating system permissions, cloud synchronization, and security software rarely alter heading styles directly, but they can prevent calibre from reading the newest source or writing a complete output file.

4.1 Move the Test Files to Local Storage

Place the temporary DOCX, debug folder, and test output in an ordinary local folder where your user account has full read and write access. Avoid converting a file while Word is still saving it, while a cloud service is synchronizing it, or while another application has it locked.

On Windows, macOS, and Linux, confirm that the file is fully available offline rather than represented by a cloud placeholder. Use a short, ordinary file name during diagnosis. If the local copy converts correctly, investigate sync conflicts or stale revisions instead of changing calibre's formatting options.

4.2 Test the Destination Device Separately

Do not involve USB transfer, email delivery, Content server access, or send-to-device conversion until the local output is correct. First open the exact converted file in calibre's viewer. Then transfer that same file to the device without reconverting it through another route.

If the local EPUB is correct but the device copy is wrong, confirm which format was actually sent. An automatic device conversion may have produced a different format with different rules. Also consider whether the device supports the CSS, image dimensions, note behavior, and output format used by the book.

Success looks like this: the same file is correct locally and on the destination device. If it is correct locally but not on the device, stop editing the DOCX and troubleshoot the device format or rendering limitation.

5. Use Logs and Conversion Debug Output

Logs are useful when calibre reports an error, silently drops content, or produces HTML that differs from the DOCX structure. They are less useful when the only difference is normal reflow between screens.

5.1 Read the Conversion Job Details

Open calibre's Jobs panel after conversion and double-click the completed job to view its detailed log. Look for warnings about unreadable resources, unsupported content, missing images, malformed links, or failed processing. Copy the log before starting another test if you may need to compare results.

A completed job does not guarantee perfect formatting, but the absence of errors helps distinguish source styling problems from a conversion failure. Focus on warnings connected to the missing or malformed content rather than unrelated informational lines.

5.2 Generate Conversion Debug Output

Use the conversion dialog's debug option and select an empty folder. calibre can save intermediate stages of the conversion pipeline, including the HTML first produced from the DOCX and the later processed content.

Open the input-stage files in a browser or text editor and locate a failing heading, image, or note. If it is already wrong in the input stage, repair the DOCX source. If it is correct there but changes later, inspect structure detection, style transformations, filtered properties, or output settings.

This provides a practical stopping rule. Do not keep rewriting the DOCX when the input-stage HTML is correct. Conversely, do not keep adding output CSS when the source has already been interpreted incorrectly.

5.3 Use Command-Line Debugging Only When Needed

Most users will not need calibre-debug or ebook-convert from a terminal. The graphical conversion log and debug pipeline are usually sufficient. Command-line tools become useful when reproducing a consistent error, collecting verbose output, or working with support to isolate a bug.

If you use a command-line test, run it against a copied DOCX and write the result to a temporary folder. Keep the graphical test and command-line test configured as similarly as possible. Otherwise, differences in options can make the comparison misleading.

6. Run a Clean Temporary Test Before Reinstalling

Reinstalling calibre rarely repairs one document's styles because the formatting information comes primarily from the DOCX and the conversion configuration. A controlled temporary test is faster and preserves your library.

1. Create a short DOCX with a title, Heading 1, Heading 2, body paragraphs, one inline image, a caption, and a real footnote.
2. Save it locally as a new DOCX.
3. Add it to calibre as a separate test book.
4. Convert it to EPUB with default settings.
5. Inspect headings, the table of contents, spacing, image placement, and note links.
6. Repeat the conversion after changing only one relevant option.

If this clean document works, calibre's basic DOCX conversion is functioning. Return to the affected source and move content into a clean document in sections until the defect appears. The last added section is likely to contain the problematic formatting or object.

If the clean document fails in exactly the same way, test with optional plugins and custom conversion rules disabled. Also test the output in another viewer. Reinstall only after a clean source, default conversion, local folder, and second viewer all reproduce a program-level failure.

6.1 Clean Up the Converted Book With Edit Book

When the conversion is structurally sound but needs a few finishing corrections, convert to EPUB or AZW3 and open Edit book. calibre's editor exposes the HTML, stylesheets, and images inside supported e-book formats.

Create a checkpoint before making broad changes. Use the live preview to inspect affected sections and Check Book to locate broken references or other detectable problems. Suitable finishing tasks include removing an accidental empty paragraph, adjusting a caption class, correcting one heading level, repairing a ToC entry, or reducing an image that exceeds the reading area.

Edit book should not become a substitute for repairing a DOCX that will be converted repeatedly. If every chapter has the same spacing problem, fix the source style or conversion rule. If only two captions need special treatment in the final edition, post-conversion cleanup may be the more efficient choice.

Success looks like this: Check Book reports no relevant unresolved problems, the live preview is correct, and the saved book also works in the actual destination reader.

7. Quick Fix Checklist

• Make a copy of the DOCX and create a short representative test.
• Use Heading 1, Heading 2, and other real Word styles.
• Replace tabs, repeated spaces, and empty paragraphs with style settings.
• Verify headings in Word's outline or Navigation pane.
• Insert images inline and place captions in separate styled paragraphs.
• Create footnotes and endnotes with Word's built-in note tools.
• Accept or reject tracked changes and remove unwanted comments.
• Save a fresh DOCX locally before adding it to calibre.
• Confirm calibre is converting the newest DOCX copy.
• Test an EPUB conversion with default Look and feel settings.
• Choose a sensible output profile without expecting fixed pagination.
• Inspect the Jobs log and conversion debug pipeline when content disappears.
• Compare the output in calibre's viewer and the destination reader.
• Use Edit book for limited final cleanup, not widespread source defects.
• Stop changing settings as soon as structure and navigation work correctly.

8. Frequently Asked Questions

8.1 Why Do My Word Headings Become Ordinary Text?

The headings may have been formatted manually with bold text and a larger font instead of a Heading style. Apply Word's built-in Heading 1, Heading 2, and Heading 3 styles, verify the outline, save a new DOCX, and convert it again.

8.2 Why Does calibre Add Too Much Space Between Paragraphs?

The DOCX may contain empty paragraphs, direct paragraph formatting, or conflicting style definitions. Show formatting marks in Word, remove unnecessary blank paragraphs and tabs, then define spacing in the body paragraph style. Also test with calibre's custom CSS and style transformations disabled.

8.3 Why Are Images Moved or Separated From Their Captions?

Floating images, text wrapping, grouped objects, and text-box captions do not map reliably to reflowable e-books. Change important images to inline placement and put each caption in a normal paragraph next to its image. Some separation can still occur when a reader changes page boundaries.

8.4 Can calibre Preserve the Exact DOCX Page Layout?

Not in a reflowable format such as EPUB. The reader controls screen dimensions, type size, margins, and pagination. Use PDF when a fixed page is essential, while recognizing that fixed pages can be uncomfortable on small screens. For EPUB, judge success by structure, readability, navigation, and image order rather than identical pages.

8.5 Why Is the Table of Contents Missing or Incorrect?

First confirm that chapter titles use real heading styles and appear in Word's heading outline. Then review calibre's Structure detection and Table of Contents options. Start with defaults, and use calibre's ToC editor for limited manual corrections after conversion.

8.6 Should I Reinstall calibre If DOCX Conversion Is Not Working?

Usually not. First convert a small clean DOCX with default settings in a local folder. If that works, the issue is in the original document or its saved conversion options. Consider reinstalling only if a clean test consistently fails across multiple documents and viewers after optional plugins and custom settings have been ruled out.