iloveepubProcessed locally — files never leave your device
Fix My EPUB
One tool for the whole rescue job: drop in a book that won't open, opens with a broken table of contents, or shows up nameless on your shelf. A health check runs first, structural damage is repaired automatically, cover and metadata gaps you fill in yourself — and the fixed book is re-checked before you download it. All of it in your browser; the file never leaves your device.
01Choose your EPUB
How to fix a broken EPUB
Drop in the broken book
Select any .epub file — one that will not open, shows a broken table of contents, or appears nameless on your shelf. Nothing is uploaded.
Read the health report
The diagnosis lists every check — container, mimetype, references, TOC, cover, metadata — marked healthy, fixable, or beyond repair.
Apply the fixes
Structural repairs run automatically. If the cover or title/author metadata is missing, add them in the same step — both are optional.
Download the verified book
The fixed EPUB is re-checked before download. You get a repair report and an all-clear, not just a blind success message.
What the health check looks at
An EPUB is a ZIP archive with strict rules about its own plumbing, and almost every "won't open" case is a plumbing problem rather than damaged text. The diagnosis inspects that plumbing layer by layer. It starts with the archive itself, then META-INF/container.xml — the shipping label that tells a reader where the package document lives. It verifies the mimetype entry: it must be the very first record in the archive, stored uncompressed, with exactly the right content string. Many repackaging tools get this wrong, and several readers refuse the book outright because of it.
Deeper in, the check compares the manifest against the actual files in the archive: every referenced document and image must exist, every spine entry must point at a declared manifest item, and the navigation documents (the EPUB 3 nav and the EPUB 2 NCX) must be present and resolvable. Finally it looks at the two things readers display on your shelf: the cover image declaration and the title and author metadata. Each finding is reported as healthy, auto-fixable, or something only you can supply.
Why EPUB files break in the first place
Most broken books are born broken by their production tool. Exporters with buggy ZIP writers put the mimetype entry in the wrong position or compress it; scrapers and homemade converters reference files they never embedded; some writing apps emit a manifest that points at paths with the wrong case on case-sensitive readers.
The rest break in transit or in storage: a download that stopped halfway, a cloud-sync conflict that merged two copies of an archive, a zip tool that "helpfully" re-packaged the book with different compression, or a manual rename that confused the container. None of this touches the actual text — which is why a structural repair, done right, restores the book without re-encoding a single chapter.
The diagnose → fix → re-check loop
A repair you cannot verify is just a guess. Desktop tools hand you a fixed file and trust you to try it; upload-based converters show a success page even when the book still fails validation. This tool closes the loop: after the structural fixes are applied, the exact same health check runs again on the result, inside your browser, before the download appears. If anything is still missing — say, you skipped the optional cover — the report says so plainly instead of declaring victory.
The re-check is also what keeps the tool honest about its limits. When a book contains damage that no browser-side repair can address, the diagnosis says cannot-be-fixed up front, before you spend a minute on it, rather than failing at the end of a fake progress bar.
What can and cannot be fixed — honestly
Fixable automatically: a missing or malformed container.xml, a mimetype entry in the wrong place or compressed, manifest items pointing at files that do not exist, spine entries referencing undeclared items, a missing EPUB 3 nav document or EPUB 2 NCX, and stale links left behind by earlier edits. Each of these has a deterministic repair that this tool applies and then verifies.
Fixable by you, with the tool doing the writing: a missing cover (upload any JPG, PNG, WebP or GIF) and missing title or author metadata (type them in; the tool writes proper Dublin Core elements).
Not fixable here, and the tool will say so: an archive that is not a valid ZIP at all, a book with no readable chapters inside, and files locked with DRM. DRM-protected books cannot be processed — removing it is neither possible in a browser nor legal in most places, and no honest tool will claim otherwise.
Why your shelf shows a filename or a blank cover
Library apps read exactly two things when they lay out a shelf: the cover declaration in the package metadata, and the dc:title / dc:creator elements. When either is missing, the reader improvises — filename instead of title, a generic gray tile instead of a cover — and people naturally assume the book file is corrupt. Usually it was simply published incomplete, or an earlier converter stripped the metadata.
That is why cover and metadata are part of this tool's diagnosis rather than a separate chore: the health check names the gap, you supply the missing piece, and the re-check confirms the shelf will now render the book properly.
Why repairing locally matters
An EPUB you care about is a personal artifact. It can carry your annotation anchors, your reading progress markers, purchase identifiers, and in the metadata, whatever the previous owner or tool wrote there. Upload-based repair services take the whole file to a server you cannot audit, and their privacy policies describe retention windows, not impossibility.
This page has no upload path at all — not a compressed fallback, not a queue, not a temporary copy. The file is opened with the same browser APIs a reader would use, modified in memory, and handed back to you. You can watch the network tab while it works: the only requests are the page itself.
Fix My EPUB vs. the single-purpose tools
This site also offers each repair on its own — Repair EPUB, Fix EPUB TOC, Metadata Editor, Change Cover — and they stay available. The difference is intent. Those pages each assume you already know what is wrong. Fix My EPUB assumes the opposite: you have one broken book and no diagnosis. It inspects everything, applies every automatic fix in one pass, prompts for the two fixes that need human input, and finishes with a verified, shelf-ready file.
Books up to 120 MB are supported, with a gentle warning past 100 MB because repairs happen in memory. The output is a standard EPUB that opens in Calibre, Thorium, Kobo, Kindle (via Send to Kindle) and every reader that accepted the original.
Frequently asked questions
My EPUB won't open at all. Can this fix it?
Often, yes — if the damage is structural. A missing container.xml, a misplaced mimetype entry, or dead internal references all prevent readers from opening a book while leaving the text intact. The health check tells you within seconds which case you are in, and if the archive itself is not valid ZIP or contains no readable chapters, it says so instead of pretending.
What is the difference between Fix My EPUB and Repair EPUB?
Repair EPUB applies structural repairs and hands you the file. Fix My EPUB is the full task: it starts with a health check across eight areas including cover and metadata, applies the structural repairs, lets you fill the gaps only a human can fill, and then automatically re-checks the result before you download it.
Is my book uploaded anywhere?
No. The diagnosis, the repairs and the re-check all run in your browser using the same parsing a desktop reader would do. There is no upload path in this page at all — you can verify it in your browser's network tab while the tool works.
Can you remove DRM from my book?
No, and no honest browser tool can. DRM-protected files are encrypted by design; the diagnosis will report them as unprocessable. This tool only repairs books you can legitimately open and modify.
Why does my reader show the filename instead of the book title?
The package metadata is missing its title or creator element, so your reader falls back to the file name. The health check flags this, and you can type the correct title and author before fixing — the tool writes proper Dublin Core metadata elements.
The table of contents is wrong. Will this rebuild it?
Yes, if the TOC document is missing entirely, it is rebuilt from the book's actual chapter list. If an existing TOC contains dead links to removed files, they are cleaned out. The Fix EPUB TOC tool remains available for rebuilding a TOC you simply dislike.
How large a book can it handle?
Files up to 120 MB are accepted, with a performance note above 100 MB because everything happens in memory. Typical novels finish the whole diagnose-fix-recheck loop in a few seconds.
Will fixing re-encode or degrade my book?
No. Chapters, images and stylesheets are moved through the archive byte-for-byte; only the broken structural files are rewritten and the archive is repacked with the mimetype entry restored to its required first position.
The diagnosis says my book cannot be fixed. What now?
That verdict appears when the file is not a valid ZIP archive, contains no readable chapters, or is DRM-protected. For a truncated download, fetch the file again from the original source. For DRM, use the store or library the book was bought from. In all three cases no data leaves your device in reaching that answer.