Your EPUB Got Rejected by the Store? Fix the Errors It Actually Names

Your EPUB Got Rejected by the Store? Fix the Errors It Actually Names

Don't send the same file back and hope the store changes its mind — the rejection points at a problem inside the file, and the file is the thing that has to change. Open the rejected EPUB in the fixer, read the four problems it names, rebuild the contents, correct the book details, and check the links once more before you send it back.

Try the toolFix EPUB Chapters & TOC: split chapters, rebuild contents, edit metadataFix just the parts of an EPUB you need right now, without learning a new ebook editor.loopyti.com

One caution before you start: if you submit the untouched file again, it will be rejected again for the same reason. The rebuild now shows its work — read the "Fixed automatically" and "Entries removed" lines under the new contents list before you download.

How to fix the named errors in three steps

  1. Open the rejected EPUB. Use the "Open an EPUB file" button, or drop the file anywhere on the page. The page says "Open one EPUB file up to 100MB. DRM-protected files are not opened." — but the header at the top of the same page says "EPUB · up to 50 MB · up to 500 files". The two limits disagree, and my test file was only 3,208 bytes, so I can't tell you which one actually applies.
  2. Read the check list, rebuild the contents, and check the book details. The tool opens the file and shows its structure: the files inside, the chapters, the contents list, the book information, and a check list of problems. Click "Rebuild contents" to rebuild the contents from the chapters that are actually there, and check the book details — title, author, language, and publisher are all editable fields, and "Apply book information" answers with "Applied the book information." The split buttons stayed greyed out in my test because each chapter already has exactly one heading, which is the normal state for a well-split book.
  3. Check the links, then download — and read the header first. The link check runs automatically when the file opens; there is no separate "check links" button to press, and the list is recalculated after a rebuild. "Download EPUB" answers with "Created broken-sample-fixed.epub." But look at the header before you trust the download: if it still shows an error count, the file still has a problem in it.

Does it actually work? I tested it

I built a fake book, "A Field Guide to Imaginary Harbors" by Marin Solace, published by Harborlight Press (fictional) — every name in it is made up — and planted three errors in it: the mimetype entry compressed instead of stored, chapter 4 present in the file and the reading order but missing from the manifest, and a contents entry pointing at a chapter file that doesn't exist. Tested 2026-11-14, and re-tested the same day after the tool was fixed (the first version of this post described the old behaviour — see below).

The moment the file opened, the header read Files 9, Chapters 4, Errors 2, To check 1. The file list showed nine files, with ch4.xhtml the only one marked Type "—" and Role "Not listed". The chapter list showed Chapter 1, 2, 3, and 5 — no Chapter 4, even though the file is really there.

![EPUB fix tool file list showing nine files with ch4 marked not listed and error counts two errors and one to check](/blog/media/2026/11/9988925c-9f3e-46e5-84e7-88fc69ad6c4a.png

The check list named four problems, in the tool's own words:

  • "Fixed on download" — "The mimetype entry's position, compression or content breaks the rule" — mimetype
  • "Error" — "Reading order refers to an unknown item ID" — ch4
  • "Check" — "File not listed in the package" — OEBPS/ch4.xhtml
  • "Error" — "Contents entry points to a missing file" — "OEBPS/nav.xhtml → ch6.xhtml"

Under the list it says: "The file is packed again with mimetype as the first, uncompressed entry. Keep your original file as it is." And a warning banner reads: "Errors remaining: 2. You can still download, but readers may not open the file."

![EPUB fix tool check list naming four problems including the mimetype rule, the unknown ch4 item, and the contents link to a missing file](/blog/media/2026/11/1e1a49a0-a8dc-4dab-b3e9-de6f3aa59f29.png

The contents list showed the damage plainly: Chapter 1 through 5, plus "Appendix — Glossary of Harbors" struck through and marked "File missing", "ch6.xhtml".

![Table of contents before fixing showing the appendix entry struck through as file missing](/blog/media/2026/11/d2774b50-f00f-4817-b088-b1578cfaa947.png

I clicked "Rebuild contents" and it answered "Rebuilt the contents with 5 entries." The contents list now has five entries — Chapter 1, 2, 3, 4, and 5 — the broken appendix entry was gone, the header dropped from Errors 2 to Errors 0, and the check list now reads "No broken internal links or missing files." Under the list, the tool shows its work: under "Fixed automatically" — "The reading order points to ch4, which was not in the file list. OEBPS/ch4.xhtml is in the package, so it was added to the list." — and under "Entries removed from the contents because their file is missing" — "Appendix — Glossary of Harbors → ch6.xhtml".

![Rebuilt table of contents with five entries including Chapter 4, with the automatic fix and the removed appendix entry listed below](/blog/media/2026/11/186383ba-6d17-4ca5-84b2-2bec8cca903a.png

Chapter 4 — Troubleshooting is still there. The rebuild builds the contents from the chapters that are actually there, and ch4 is a real file — so this time the chapter stayed, and only the broken appendix entry went. (An earlier version of this post said the rebuild deleted Chapter 4 along with the broken entry — that was true before the fix, and it is fixed now.)

The book details were untouched by all of this: Title "A Field Guide to Imaginary Harbors", Author "Marin Solace", Language "en", Publisher "Harborlight Press (fictional)", Description empty. I applied them unchanged and got "Applied the book information." There is no cover image — "There is no cover image." — and the "Download cover" button is disabled.

Then I downloaded. "Created broken-sample-fixed.epub." The file is 3,200 bytes. I opened the downloaded file as a zip and checked it directly: the mimetype entry is first, stored uncompressed, and reads "application/epub+zip" — so the "Fixed on download" claim is real. The manifest now lists ch4, the spine still refers to ch4, and the contents links are ch1, ch2, ch3, ch4, and ch5 — the broken link is gone because its entry was deleted, and the chapter that was missing from the manifest is now in it.

And the header after the download reads Files 9, Chapters 5, Errors 0, To check 0, and the check list says "No broken internal links or missing files." There is no "errors remaining" warning any more.

![Check and download section showing no broken internal links or missing files after the EPUB was rebuilt](/blog/media/2026/11/cc4978cc-c692-4f90-ba2c-e138d0c90c72.png

So the honest score for my three planted errors: the mimetype problem was really fixed, the broken contents link was fixed by deleting its entry, and the manifest gap was repaired too — ch4 is in the manifest now. All three are fixed.

Why a list and its files have to agree

An EPUB is a zip of small files plus a contents list — the manifest and the table of contents — that says which files exist and in what order they are read. When the list and the files disagree — the list points at a file that isn't there, or a file exists that the list never mentions — the store's checker stops trusting the book. Fixing it is not rewriting the story; it is making the list match the files. The mimetype rule is the one exception here: that entry has to sit first and uncompressed, and this tool repacks it correctly on download, so that part is handled for you.

What about the other ways?

Fix it by hand in Calibre or Sigil works — you can add a chapter to the manifest yourself. But you have to install and learn an ebook editor, and editing the manifest and the contents list by hand is exactly how they drift apart again. This tool now does the manifest repair itself, when the file is actually in the package.

Run a validator like epubcheck and fix the list it gives you gives you the most accurate error list. What it doesn't give you is the fix: you still have to unzip the book, edit the files, and zip it back up — and repacking a zip by hand is how the mimetype rule gets broken in the first place.

Export the book again from the original source file is the cleanest fix when you still have the source. But if all you have is the rejected file, or the export produces the same error again, there is nothing to re-export from.

This tool's place is the middle case: the store named an error in the contents or the book details, and you want to fix that part in the browser without installing anything. One thing to be clear about: this tool is not a validator. It doesn't run the store's checks and hand you a pass or fail. It shows you where the list and the files disagree, and lets you fix the contents and the metadata. The error list it shows is its own check, not the store's verdict.

What it won't do — read this first

  • It now fixes a chapter missing from the manifest — when the file is actually in the package, the rebuild adds it to the list and says so. (An earlier version of this post said there was no way to add it — that was true before the fix, and it is fixed now.) What I did not test: a chapter the reading order points at whose file is not in the package at all. The fix note says the rebuild removes the spine reference in that case — so check the "Fixed automatically" lines for your own file.
  • Rebuilding the contents now keeps real chapters and shows what it removed. The rebuild kept Chapter 4 and listed the removed appendix entry by name. (Before the fix it deleted Chapter 4 silently.) Read the rebuilt list anyway — it is the last chance to catch a missing chapter before you download.
  • The mimetype diagnosis doesn't say what is actually wrong. The check list only says the entry's "position, compression or content breaks the rule" — it never says which of the three it is. In my file the problem was compression, and the download really did fix it (verified in the downloaded zip: first entry, stored, "application/epub+zip"), but the tool itself never tells you that.
  • A download is not a pass. In my test the rebuild cleared every error, so no warning appeared — the header read Errors 0. The old version let you download a file that still had errors, and I did not re-test that path, so if your header still shows errors after a rebuild, do not treat the download as fixed.
  • It won't open DRM-protected books, and it won't rewrite body text. DRM files are not opened at all, and this is not a text editor — chapter structure, contents, metadata, and cover are the whole range. A typo in the body stays a typo.
  • Your file stays in your browser. Nothing is sent to a server, and closing the window clears what you entered. Keep a separate copy of the original before you edit — the tool says so itself, and my test left the original file untouched.

The one line to remember

Fix what the checker named — then read the contents list before you resubmit.

Try the toolFix EPUB Chapters & TOC: split chapters, rebuild contents, edit metadataFix just the parts of an EPUB you need right now, without learning a new ebook editor.loopyti.com

FAQ

Does it remove DRM? No. DRM-protected files are not opened at all — the page says so, and there is no way around it in the tool.

Can I rewrite the body text? No. It edits chapter structure, the contents list, metadata, and the cover. Body text needs an ebook editor.