[Conversion Plugin] KFX Input

[tomsem]

Originally Posted by jhowell

In the notebook with the new shader you have for each color a short line along with a longer line labeled "1" and "12". What do these represent?

Can you share what these notebook pages actually look like on your device or when exported?

The number of 'strokes' in any given square. The Shader tool adds to apparent saturation according to number of times you draw over a region with it. I don't know how many 'overlays' it takes to get to 100% but it kinda looks like it maxes out at 9 or 10.

[jhowell]

Originally Posted by tomsem

The number of 'strokes' in any given square. The Shader tool adds to apparent saturation according to number of times you draw over a region with it. I don't know how many 'overlays' it takes to get to 100% but it kinda looks like it maxes out at 9 or 10.

Thanks.

I will release an update to this plugin with support for Scribe Colorsoft notebooks.

[jhowell]

Version 2.28.0 - 06 Jan 2026

Handle conversion of notebooks generated by Scribe Colorsoft firmware 5.19.1. This includes the shader brush, stroke colors, and text generated by handwriting recognition. Recognized text will be present as a transparent overlay in the SVG files produced.

[tomsem]

Originally Posted by jhowell

Version 2.28.0 - 06 Jan 2026

Handle conversion of notebooks generated by Scribe Colorsoft firmware 5.19.1. This includes the shader brush, stroke colors, and text generated by handwriting recognition. Recognized text will be present as a transparent overlay in the SVG files produced.

Very cool and thanks! It should work on those created with new B&W Scribe as well? Looking forward to trying it.

I have been curious about how they would handle notebook sync between 10.1" Scribe and 11" Scribe.

There is a Book and Notebook options setting called Notebook Scaling: 'Scale the notebook page to fit to screen. Close and reopen your notebook to see the changes'. I think default is ON, at least I do not remember changing it.

I have been assuming that scaling to fit screen was a requirement. Surely it would cause issues with sync if the templates were different (different number of lines, dots etc. on the different screen sizes).

I guess I need to turn it OFF and open some notebook I crated on my 2022 Scribe to 'see the changes'?

[rooobooot]

Hi jhowell and everyone,

*First, a quick note: My English is not very strong, so this post has been edited and optimized by Grok (an AI assistant) to make it clearer and easier to read.*

**Environment information (for quick reference):**
- OS: Windows 11
- Calibre: 8.16.2
- KFX Input plugin: v2.27.1
- DeDRM: v10.0.9
- Kindle for PC: 5.16.9 (4246270033)

I am a legitimate purchaser of Kindle books in China. Unfortunately, Amazon officially exited the Chinese market (Kindle China store closed in June 2023 and the app was removed in 2024), so it is no longer possible to re-download purchased books in older AZW/AZW3 formats. This is the reason I use the well-known community tools to remove DRM from my legally bought books for personal backup and format conversion purposes only.

I have carefully read the MobileRead posting rules and fully understand that direct discussion of DRM removal is not permitted here. I mention the above only to provide necessary context about why I only have access to these newer multi-container KFX files.

The KFX Input plugin has worked wonderfully for me – I have successfully imported and converted over 100 of my Kindle books. Only three very large bundled editions remain problematic. This post concerns one of them: the Chinese bundled edition of "A Song of Ice and Fire" volumes 1-5 (full 15 books in one ASIN), title "冰与火之歌1-5卷（全15册）", ASIN B00KYEFXK4.

DeDRM successfully decrypts the file, but KFX Input fails during the packaging step with the error:

"Book appears to have KFX containers from multiple books. (duplicate fragments)"

There are numerous warnings about duplicate fragment keys (mostly image resources), fragment ids with multiple types $417/$597, containers missing from entity map, and unreferenced fragments.

Manual unpacking of the KFX-ZIP and importing individual containers allows them to be added to the library as KFX, but subsequent conversion to EPUB/PDF/etc. still fails with KeyError('$165').

To make testing and reproduction as easy as possible, I have uploaded the following files to a public GitHub repository:

- Original .kfx file downloaded from my Kindle device (still containing DRM):
https://github.com/rooobooot/kindle-...B00KYEFXK4.kfx

- Decrypted KFX-ZIP file produced by the tools:
[SNIP: link to decoded in-copyright file]

- Full debug log obtained with calibre-debug -g:
https://github.com/rooobooot/kindle-...b/main/log.txt

Is there any known workaround for such extremely large multi-volume bundled KFX books, or would it be possible to add an option in a future version to ignore or better tolerate duplicate fragments?

Thank you very much, jhowell, for your outstanding work on the KFX Input plugin – it has already helped me preserve most of my library!

Best regards,
rooobooot

[jhowell]

Originally Posted by tomsem

It should work on those created with new B&W Scribe as well?

You will have to let me know how it works. I am happy to look at any further notebooks that you care to provide.

[jhowell]

Originally Posted by rooobooot

DeDRM successfully decrypts the file, but KFX Input fails during the packaging step with the error:

"Book appears to have KFX containers from multiple books. (duplicate fragments)"

I have seen this error occasionally. It is caused by files from different books (or different revisions of the same book) being present in the folders associated with a book.

Normally I would suggest removing the book from your device and re-downloading it from Amazon in order to correct the problem. But that will not be possible in your situation.

You should be able to make it work using a more slightly more complex procedure instead. You can rename your .kfx-zip file to have a .zip extension, then delete the extra files that are causing a problem from it using any tool for manipulating zip files, and finally rename it back to .kfx-zip. You should then be able to convert the resulting file using this plugin.

This message from your log indicates the files that need to be removed:

"Found containers missing from entity map: CR!RZ4V5Q8ZEN0HXC78C3TY5TT27VSH, CR!VFQM6NX6850X5EPNWF2V42Q8NHF4, CR!VGQBHT5GZ93P5EJK6J4NXV5GEC8S, CR!YHWQ50SDWH5BS5Z9SD48J8HA8KMM"

Let me know if you need further assistance.

[tomsem]

Originally Posted by jhowell

You will have to let me know how it works. I am happy to look at any further notebooks that you care to provide.

Yes, I don't have the new B&W but can see what happens with the 2022 Scribe when I sync over a notebook created on the KSC. When I get a chance...

[rooobooot]

Originally Posted by jhowell

I have seen this error occasionally. It is caused by files from different books (or different revisions of the same book) being present in the folders associated with a book.

Normally I would suggest removing the book from your device and re-downloading it from Amazon in order to correct the problem. But that will not be possible in your situation.

You should be able to make it work using a more slightly more complex procedure instead. You can rename your .kfx-zip file to have a .zip extension, then delete the extra files that are causing a problem from it using any tool for manipulating zip files, and finally rename it back to .kfx-zip. You should then be able to convert the resulting file using this plugin.

This message from your log indicates the files that need to be removed:

"Found containers missing from entity map: CR!RZ4V5Q8ZEN0HXC78C3TY5TT27VSH, CR!VFQM6NX6850X5EPNWF2V42Q8NHF4, CR!VGQBHT5GZ93P5EJK6J4NXV5GEC8S, CR!YHWQ50SDWH5BS5Z9SD48J8HA8KMM"

Let me know if you need further assistance.

Hi jhowell,

Thank you so much for your quick and detailed response – your suggestion worked perfectly!

I followed your instructions: renamed the decrypted .kfx-zip file to .zip, opened it with a ZIP tool (I used 7-Zip), and deleted the four containers listed in the log:

CR!RZ4V5Q8ZEN0HXC78C3TY5TT27VSH
CR!VFQM6NX6850X5EPNWF2V42Q8NHF4
CR!VGQBHT5GZ93P5EJK6J4NXV5GEC8S
CR!YHWQ50SDWH5BS5Z9SD48J8HA8KMM

After removing those files and renaming it back to .kfx-zip, the book imported and converted without any issues. Even better – I applied the same method to the other two large bundled books that were giving me problems, and they all work perfectly now. My entire library is finally preserved!

I vaguely remember that these bundled editions may have received updates or revisions from the publisher over the years, so those extra containers were probably leftovers from older versions – that would explain why they were causing the conflict.

I've removed the uploaded files from my GitHub repository for privacy, but for anyone who encounters a similar issue in the future, the key part of the log to look for is this line:

"Found containers missing from entity map:[list of container IDs]"

Deleting exactly those files from the decrypted KFX-ZIP resolves the "duplicate fragments" problem.

Once again, thank you very much for your outstanding work on the KFX Input plugin and for taking the time to help. It has been invaluable in saving hundreds of my books!

Best regards,
rooobooot

[artful]

Hello everyone.
I got this error message converting an ebook

calibre, version 8.16.2
ERRORE: KFX conversion failed: <p><b>Cannot convert Isabella Nagg e il vaso di basilico (Italian Edition)</b><br/><br/>Exception('locate_offset failed to find offset 0 (remaining=0, split_after=True) in b\'&lt;div&gt;&lt;img src="image_rsrc2H5.jpg" alt="" style="-kfx-render: inline; -kfx-style-name: s7U; box-sizing: content-box; width: 20.752%%"/&gt;&lt;span/&gt;&lt;/div&gt;\'')</p>

DeDRM: 10.0.15
KFX Input 2.28.0

Is there a way to fix it?
Thanks in advance

[jhowell]

Originally Posted by artful

I got this error message converting an ebook
...

I have encountered that error occasionally. Try selecting "Allow conversion to proceed even if errors are detected" in the KFX Input tab of the conversion dialog. That should allow conversion to complete despite this problem but there may be some minor formatting issues in the resulting EPUB.

[artful]

Originally Posted by jhowell

I have encountered that error occasionally. Try selecting "Allow conversion to proceed even if errors are detected" in the KFX Input tab of the conversion dialog. That should allow conversion to complete despite this problem but there may be some minor formatting issues in the resulting EPUB.

Thanks for your reply.
I followed your advice and the conversion was successful. The book is full of small images embedded in the text, and that may be the problem. In any case, it's readable without any problems. Thanks again.