Start with one question
Is the problem in the conversion, or in the book? Open the original in the EPUB Reader or another app. If the book already looks wrong there, converting it will not help; repair the book first. If it looks right there, the problem is in the target format or its settings.
Symptoms and causes
| Symptom | Likely cause | What to do |
|---|---|---|
| Letters like “é” instead of “é” | Text opened with the wrong encoding | Open the TXT as UTF-8. The output is UTF-8; the viewer is guessing wrong. |
| Squares instead of Arabic, Hebrew or CJK | No system font for the script | Install a font that covers it, or switch viewer. |
| Arabic letters appear disconnected or reversed | Viewer does not support complex text layout | Use a viewer that does; for PDF, use the browser print route which does. |
| Some chapters are absent | Chapters exist in the archive but not in the reading order | Validate; use the repair tool to re-link chapters that navigation still lists. |
| Conversion fails immediately | Not a ZIP, truncated, or unsafe archive | Download again. Files with unsafe paths or extreme compression are refused on purpose. |
| Output is gibberish | DRM-encrypted content | These tools cannot decrypt protected books. |
| Images missing in PDF | Image path not found in the archive, or image over 4 MB | Validate for missing resources; shrink very large images. |
| Tab crashes or freezes | Large book exceeds device memory | Use a desktop browser, close other tabs, or split the book. |
Strange characters like in text | Double-escaped entities in the book | Run the repair tool, or find and replace after conversion. |
| Headings missing structure in TXT | Plain text has no heading levels | Convert to Markdown instead. |
Diagnose in this order
- Validate. The EPUB Validator takes seconds and removes whole classes of guessing.
- Repair. If it reports errors, run the repair tool and convert the repaired copy.
- Try another target. If TXT is fine but PDF is wrong, the book’s content is intact and the issue is layout or fonts.
- Try another device. Memory problems are device problems. The same book often converts on a computer.
Large files
EPUB tools in the browser hold the whole archive and its expanded contents in memory. A book with hundreds of high-resolution images can need many times its file size. The tools refuse files over 100 MB, archives with more than 8,000 entries, or ones that would expand beyond 500 MB, so a problem cannot become a frozen browser. If you hit a limit with a legitimate book, shrink the images with an image tool or split the book.
What these tools do not try to do
- Remove DRM.
- Convert fixed-layout books (comics, picture books) into reflowable text.
- Recover text from scanned image-only books; that needs OCR.
- Guarantee identical typography across formats; each format sets its own limits.
When something fails and the message is not helpful, check the file with W3C EPUBCheck. It reports the exact rule and file, and that information is what you need to fix the book or to ask for help.
Reading the tool’s messages
| Message | Meaning | Next step |
|---|---|---|
| This is not an EPUB/ZIP file | The file does not start with a ZIP signature | Check it is really an EPUB; re-download. A web page saved as .epub is a common cause. |
| The archive is damaged or uses an unsupported ZIP feature | The ZIP structure could not be read | Re-download. Encrypted ZIP entries are not supported. |
| Contains an unsafe file path | An entry tries to escape the book’s folder (for example ../) | The file was refused on purpose. Obtain a clean copy. |
| Would expand to an unsafe size | Possible decompression bomb | Refused on purpose. If the book is legitimate, split or recompress its images with another tool. |
| No readable package document | container.xml or the package file is missing or broken | Run the repair tool. |
| Declares encrypted or obfuscated resources | DRM or font obfuscation | These tools do not process such books. |
When to stop trying
If the validator reports a damaged archive, the repair tool finds nothing to repair, and the book was downloaded from a source you cannot re-download from, the file is probably beyond rescue. A truncated ZIP lost its central directory, and no amount of tooling brings the missing data back. Ask the seller or publisher for a new file, or restore from a backup.
Preventing the problems
- Do not unzip, edit and re-zip books with a general archiver. Use tools that write EPUB correctly.
- Keep an original copy before running any tool on a book.
- Validate new books when you receive them, not months later when you need them.
- Keep metadata tidy; apps that rebuild a library from files rely on it.