I've had this idea for a year now, and the other post on this forum kickstarted it again. Claude is incredibly helpful here and makes the development an ease. It's still in Alpha and I did a testrun this past Sunday in church. It actually held up pretty nice, but I updated it a lot since then.

The basic principle is this: Whisper converts spoken words to text -> NLLB or LibreTranslate then translates that text output to the chosen language. This all happens within seconds.

It's designed for churches, but you can use this in any project.

0.1.9 is out now. It has an auto update system, so whenever I publish a new version to GitHub (which happens quite often since it's in heavy development), you will see an update notice in the software.

There is a tie-in to vMix (live translations in your stream) and to OBS Studio.

Go try it for yourself. If you have any issues or requests, open an issue on GitHub.

A usage guide and install instructions are all found on GitHub.

https://github.com/guido-visser/live-translate

P.s. I completely understand if this post is removed, as per the forum rules. The code is generated by Claude. I didn't know that was a rule until I opened the rules again. But to be fair, I am a software developer. Claude is just so much quicker (and sometimes better) than me 😅. So I'm sorry about that, if this post is removed, so be it.