Footnotes not compiling to ODT format

Hi all,

My footnotes are not properly compiling to ODT format. In ODT format, the footnotes show up directly inline with no separation in ODT. They are properly formatted as footnotes (in the footer) in PDF, RDT, and other file formats. This also does not change even if I enable inline/inspector footnotes to endnotes.

I just upgraded from Scrivener 2 to 3.0.3 (latest version), and this occurred right away. It was working fine for me on the same project with Scrivener 2. ODT is really important since I use Scrivener with Zotero, and Zotero is only able to process ODT files. Converting from RTF doesn’t work either as it loses the proper footnote formatting when you convert from RTF-> ODT.

Any help would be greatly appreciated. If I can’t find a fix, I’ll unfortunately need to downgrade back to Scrivener 2 even though I just paid for the upgrade to 3.


I can’t reproduce this problem. I just tried exporting to ODT, opened the file in LibreOffice, and the footnotes were all there at the bottom of the pages as expected. Some things to check:

  1. Make sure you are exporting “OpenOffice (.odt)” and not to “MultiMarkdown → OpenOffice (.odt)”.

  2. What app are you opening the file in?

  3. What Compile Format are you using? There are some formats, such as APA, which places footnotes at the end of the text instead of at the bottom of the page.

  4. In Preferences > Sharing > Conversion, ensure that “Use enhanced converters” is ticked" (it should be unless you have explicitly changed it though).

All the best,

  1. I am not using MultiMarkdown
  2. LibreOffice, which ought to be the native application for ODT. I’m fairly sure it’s not an application problem since it opens other ODT files generated from Scrivener 2.5 without issue.
  3. I’m using a Chicago Essay project format, but it’s not a compile format issue. I’ve noted above that footnotes display properly using the same format settings except exported to RTF, PDF, or DOC. The way the footnotes are displayed in ODT is clearly a compile error – it does not resemble intentional formatting.
  4. I did not explicitly change this, and I just double checked and Enhanced Converters is checked.

I would be happy to attach screenshots, or perhaps send over any logging data or a sample project.

If you could zip up and attach a sample project that demonstrates the issue, that would be really helpful, since I’m not able to replicate the issue my end.

Thanks and all the best,

Thanks, Keith. I just sent a sample project to the email address.

Thanks. I’ve reproduced the issue and it seems to be down to a bug in the third-party converter we use which has been fixed in the updated version that will be embedded in 3.1.

All the best,

Thank you!

I’m only resurrecting this thread because I just bought Scriv 3.3.6 in early March 2024 (MacOS Monterey 12.7.4) and the first time I tried to export some newly-written pages to .odt, I got an error message saying that if I did this, I would lose the formatting. It suggested using .rtf instead. I didn’t really want to, for various reasons, and managed to stumble upon a workaround:

I’ve never had Word on this machine, but I told Scriv to export to the old version of it (Word 1997-2004) which has the .doc suffix, not .docx. It did so, and I was able to access my page in Open Office via the “Open with” command, then saving it again and telling it to save as an .odt rather than .doc. Once this was done, it’s stayed in .odt ever since.

So the problem has an easy enough fix, on my MacBook anyway. And since Word is a MS product, I imagine they wouldn’t deny it to their own Windows user base.

You probably got that message because you don’t have Java installed on your Mac. Scrivener has a superior ODT converter built-in, but it only uses it if you have Java. If you don’t, then it falls back to using Apple’s generator, which is very basic.

Temurin is the easiest way to install it. Note the build is not “universal” so you need to pick one depending on Intel/Silicon hardware. I would also recommend the “.pkg” download if you want a user-friendly installer.

Otherwise, I would expect the .docx export to be superior to .doc. The former is the modern version, and is likely to have better support in other word processors, plus it is a file format Scrivener supports directly rather than going through third-party conversion.


Question: what is the difference between LTS versions and non-LTS versions? I presume the LTS are stable versions… am I right?

I’ve installed 21LTS through Homebrew, but it would be good to be sure.


1 Like

Oops, that helped me notice I accidentally linked to an older version. LTS builds (or Long-Term Support) are going to be of more interest to developers than users just wanting to run Java software. These are the versions that will receive security updates for years after they would otherwise be retired, and so programmers want to use those to target their code against (to avoid constantly having to update their code, like Apple programmers do :laughing: ).

1 Like

So, as a mere end-user, I ought to uninstall 21-LTS and install 22 as that will update every time I run brew update; brew upgrade, or is that unnecessary?

1 Like

Yeah, it’s probably a good idea to stick with the latest for the sake of simplicity, and let brew keep things up to date automatically. I’m not sure how the older version streams work with brew, whether the package maintainer keeps them up to date with patches or not—I would suspect so, but I don’t know.

I just double-checked, and I’m on 22+36. If you only use it for Scrivener though, I wouldn’t stress about it, since this module we use to produce ODT is purely offline.

1 Like

It seems like you’re dealing with a different issue. The footnote issue was fixed in a subsequent release of Scrivener.

In my experience, the message that you lose some formatting is relatively innocuous. I’ve rarely had issues in the final ODT. Scrivener’s built in ODT converter works quite well.


This may be a different message than what you are thinking of. There is a more dire warning if you don’t have Java installed, and aren’t using Scrivener’s ODT conversion, but rather the Mac’s. It’s basically what you ended up seeing those years ago, because what happened back then is that if the better ODT converter failed, it would fall back to using the system converter, and that’s why you lost footnotes. The Apple converters lose a number of things some would consider critical, like images if I recall correctly.

That is… unless Apple’s converters are better these days for ODT and that is what you’re using now. I honestly haven’t tested it in ages because it’s always been one of those things you never want anyone to really use.

1 Like

Ah, I have Java installed so perhaps that’s why I haven’t seen this message!