iloveepubProcessed locally — files never leave your device
Repair EPUB
A book that throws "not a valid EPUB" usually has a broken package structure, not broken content. This tool runs a health check, rebuilds the container and metadata entries, clears dead references and hands you back a cleanly packed file — all locally in your browser, because shipping a broken book to a server helps no one. Free, no upload.
01Select your EPUB
How to repair an EPUB file
Select your broken EPUB
Open the file from your device. It is examined locally — a damaged book is exactly the kind of file you don't want to upload.
Run the health check
The structure is checked and everything fixable is repaired: container, mimetype entry, dead references, missing TOC.
Download the repaired EPUB
Save the fixed file and open it in your reader. A report shows exactly what was repaired.
Why EPUB files break
An EPUB is a ZIP with strict rules: an uncompressed mimetype entry first, a container.xml pointing at the package file, and a manifest that matches the files inside. Sloppy exporters, interrupted downloads and manual re-zipping break those rules constantly — and readers refuse the whole book for it.
The good news: the chapters themselves are almost always intact. Fixing the packaging is enough to bring the book back.
What can — and cannot — be repaired
Repairable: missing or invalid container.xml, misplaced or compressed mimetype entries, manifest and spine references to files that do not exist, missing navigation documents.
Not repairable: a ZIP archive truncated mid-download or corrupted at the byte level, and content that was never in the file. No tool can rebuild chapters that are not there — if repair fails, re-download the original file.
Frequently asked questions
What kinds of damage can this repair?
Broken packaging: invalid container.xml, wrong mimetype entry, manifest or spine references to missing files, and missing table-of-contents files. These cover the majority of "won't open" cases.
My file is truncated or half-downloaded. Can it be fixed?
No. If the ZIP data itself is cut off or corrupted at the byte level, the missing content cannot be invented. Re-download the original file instead.
Does repair change the book's content?
No. Chapters, styles, images and fonts are copied byte-for-byte. Only structural packaging is rebuilt — and a report lists every change.
The repaired file still won't open. What now?
The damage is likely at the byte level of the archive itself. Try re-downloading or re-exporting the original; a repack cannot resurrect missing bytes.
Why do readers care about the mimetype entry so much?
The spec requires it as the first, uncompressed entry — many readers check it before reading anything else. A wrong mimetype is the most common and most fixable breakage.
Is my book uploaded to a server?
Never. The entire repair runs in your browser — verifiable in the network inspector: zero uploads.