You tap an ebook and nothing happens. Or worse — the cover loads, the reader crashes on chapter two, and the table of contents points to pages that don't exist. A broken EPUB can fail in confusing ways because an EPUB is not one file but a package: HTML chapters, images, a stylesheet, and several XML files that describe how it all fits together, all zipped together. Any one of those parts can break the whole book.
This guide covers the five fixes that resolve almost every corrupted EPUB, ordered from fastest to most thorough — and explains what actually broke so you can stop it happening again.
How to Tell What Broke
Before fixing, identify the symptom. Each failure mode points to a different repair:
| Symptom | Likely cause | Fix |
|---|---|---|
| File won't open at all | Broken ZIP container or missing mimetype | Method 1 or 2 |
| Opens, then crashes mid-book | One malformed chapter file | Method 2 |
| Table of contents is empty or wrong | Damaged nav / NCX file | Method 4 |
| Cover missing, book otherwise fine | Cover image or OPF reference broken | Method 2 |
| "Invalid file" in a specific app only | Strict validation (often Kindle's converter) | Method 3 then 1 |
If you are unsure, start with validation (Method 3) — the error message usually names the exact broken file, which turns a mystery into a to-do list.
Method 1: Automatic EPUB Repair (Fastest)
Purpose-built repair tools rebuild the package structure for you: they unpack the book, repair or regenerate the container files, fix references, and repack a clean EPUB.
The workflow takes under a minute with the free EPUB repair tool:
- Open the tool and drop in the broken
.epubfile. - It unpacks, diagnoses, and rebuilds the package — repairing the container, OPF references, and file paths that trips most readers.
- Download the repaired book and open it in your reader.
Because the tool runs in your browser, nothing is uploaded — relevant if the file is an unpublished manuscript or an ARC you agreed to keep private.
Automatic repair resolves the majority of failures, because most broken EPUBs suffer from packaging errors rather than damaged content: a missing mimetype entry, wrong file paths in the OPF, or a ZIP written in the wrong order by whatever tool produced it.
Method 2: Open It as a ZIP and Fix the Part That Broke
An EPUB is a ZIP archive with the extension renamed. When you know exactly which file is damaged, manual surgery is faster than any tool:
- Copy the file and change the copy's extension from
.epubto.zip. - Extract it. If extraction fails, the ZIP container itself is damaged — go back to Method 1.
- Inspect the parts:
mimetype— must be the first, uncompressed entry, containing exactlyapplication/epub+zipMETA-INF/container.xml— points to the OPF file- The
.opffile — the manifest listing every chapter, image, and stylesheet - Chapter HTML files — open them in a browser to find the one that breaks
- Fix the broken file — correct the path in the OPF, restore a deleted chapter, or replace a corrupt image.
- Repack correctly: zip the
mimetypefile first (stored, not compressed), then everything else, and rename back to.epub.
The repacking detail matters — getting it wrong produces a book that works in some readers and fails in others. If your ZIP tool fights you on this, that is exactly what Method 1 automates.
Method 3: Validate First When You Don't Know What Broke
Guessing is slow. EPUBCheck is the official validation engine that reads your file and reports every standards violation with file names and line numbers.
Run the EPUB checker to get a report before and after any repair:
- Before: identifies the actual failure — often something as small as one unclosed tag in chapter 7 or one image missing from the manifest
- After: confirms your fix is complete rather than hoping the book "seems fine now"
Validation is also the go-to fix for the Kindle-specific case: a book that opens in every app but fails Send-to-Kindle. Kindle's converter is stricter than most readers, and the validation report tells you exactly which violation it choked on.
Method 4: Rebuild the Table of Contents
A specialized but extremely common breakage: the book opens and reads fine, but the table of contents is empty, duplicated, or jumps to the wrong chapters.
This happens because EPUB carries its navigation in one or two dedicated files (the EPUB 3 nav.xhtml and the legacy ncx), and they drift out of sync with the chapters whenever a book is edited — files renamed, chapters split, sections reordered.
You rarely need to fix those files by hand. The TOC rebuilder scans the actual chapters in the book, rebuilds the navigation from the real heading structure, and writes a clean TOC that readers understand. It also handles the mixed EPUB 2 / EPUB 3 situation where one of the two navigation files is fine and the other is stale.
Method 5: Salvage Through an Editor
When the book is mostly fine but one chapter is visibly damaged — half a page of mojibake, a missing image, a duplicated section — open it in the online EPUB editor instead:
- Open the book and browse the chapter list to locate the damaged file.
- Edit or delete the broken content directly, with live preview.
- Fix the metadata or cover while you are in there.
- Export a clean copy, then validate it (Method 3) to confirm.
This is also the right method when the "breakage" is really obsolescence — books a decade old with nonstandard internals. An editor that rewrites the package on export often modernizes the book into something every current reader accepts.
Why EPUB Files Break in the First Place
- Bad exports. Most corrupted EPUBs were born that way — a conversion tool wrote an invalid package.
- Interrupted transfers. A download or sync that stopped midway leaves a truncated ZIP.
- Editing accidents. Manual edits that renamed a file without updating the manifest.
- Storage rot. Rare, but flash storage and old CDs do corrupt files.
- Aggressive "optimization." Some tools recompress EPUBs as plain ZIPs, breaking the mimetype rule silently.
The defensive habits: validate books right after converting or editing them, keep originals, and let a repair tool rebuild the package rather than trusting whatever wrote it.
Frequently Asked Questions
Can a corrupted EPUB be fixed without the original file? Usually yes — the content (chapters, images) is typically intact even when the packaging breaks. Repair tools rebuild the container around the existing content. Truncated downloads are the exception: missing bytes at the end of the file can mean missing chapters.
Why does my EPUB open on one device but not another? Readers differ in strictness. Lenient apps skip over errors; strict ones (notably Kindle's converter) reject the whole book. Run the EPUB validation and fix every reported error, and the book will open everywhere.
Is a repaired EPUB safe to side-load to Kindle or Kobo? Yes. Repair only rebuilds the package structure; the content is unchanged. Validate after repairing if the book is going to a strict reader.
How do I fix an EPUB that has no cover? A missing cover is usually a broken reference in the OPF, not a lost image — Method 2 territory. The EPUB cover changer fixes it in one step by embedding a new cover with correct declarations.
Fix It Now
Start with the symptom table above to pick your method. When in doubt: validate, repair, validate — three minutes total with the EPUB checker and the repair tool, and nearly every broken EPUB comes back to life.