What is getting removed Load a file to see
Off, because they break something
Author, publisher, dates, rights, description and source come out every time. There is no argument for keeping them, so there is no switch.
Writes a second copy. The file on your device is not modified. Language always stays, because the spec requires it and it identifies nobody.
Sections
Also do this
Sort names are written in both the EPUB 2 and EPUB 3 dialects, because converters disagree about which to read. Any record field not shown above is cleared.
Choose a file and everything it is carrying appears here, field by field, with what happens to each one.
The record, field by field
Open What is getting removed to change any of this.Book
No cover image declared.
JPEG, PNG or WebP. Tall and narrow, around 1600 by 2560, is what reading devices expect. Metadata inside the image you choose is stripped before it goes in, so a cover shot on a phone does not carry its GPS coordinates into the book.
What this tool removes, and what it cannot
An EPUB carries the name of whoever bought it, the shop that sold it and the app that opened it last. Almost none of that is in the book. It is in the file record, and it travels with the file wherever the file goes.
Book metadata
- Author, contributor, publisher
- Names a person or a company. The most common reason anyone strips an EPUB in the first place.
- Publication and modification dates
- Dates the file, which narrows down where and when it came from.
- Rights and licence text
- Frequently carries the licensee name or the email address the file was issued to.
- Description, source and relation
- Source fields point at the catalogue, store or ISBN the copy came from.
- Calibre fields
- Series, sort order and library timestamps written in by the desktop app. None of it belongs to the book.
- Sort and role refinements
- Filed under the author entry. Removing the author but leaving the sort name leaves the name.
- Subjects and tags
- Personal shelving categories. Optional, because some are genuinely part of the book record.
- Cover image
- Optional. Cover art carries its own EXIF, including the location a photograph was taken.
The archive itself
- File timestamps
- An EPUB is a zip. Every file inside records when it was written, which dates the download.
- Extended attributes
- A second, higher precision timestamp hides here. Normalising the visible date alone leaves it behind.
- Archive comments
- Free text attached to entries by whatever built the file.
Reader leftovers
- Bookmarks and reading position
- Records how far through the book you were and what you highlighted.
- Store receipts
- Apple and other stores write an account identifier into the file at download.
- Desktop clutter
- Finder and Explorer index files that got zipped in by accident.
What it cannot remove
Three fields are required by the EPUB specification. Strip them and the file stops opening in most readers, so the tool replaces them instead of deleting them and tells you it did.
- Identifier
- Every EPUB must declare one. The original is often a library UUID or a store transaction reference, so a fresh random UUID replaces it rather than nothing at all.
- Title
- Required by the spec. Kept by default, because a reading device falls back to the filename when the title is missing.
- Language
- Required by the spec, and identifies nobody.
What it does not touch
The text of the book. Every file inside the archive except the package document is copied across as raw compressed bytes, so nothing is re-encoded and no chapter is re-rendered. Metadata written inside the chapters themselves, or a name printed on the title page, stays where it is. This works on the file record, not the contents.
Where the file goes
Nowhere. The archive is unpacked, rewritten and handed back inside this browser tab. There is no upload, no server round trip and no third party script involved at any point, which is why the tool works with the network switched off.