KDEnlive offers two modes for speech-to-text generation with which to automatically create subtitles for videos. VOSK may have been available in earlier versions of KDEnlive, but I didn't know about it until recently. VOSK requires the dev-python/srt package which is not currently available in the main Gentoo repo but is available in the Guru overlay. It also requires the VOSK Python module which doesn't appear to be available in any overlay. Its homepage is https://github.com/alphacep/vosk-api The new speech-to-text mode in KDEnlive is Whisper. Whisper requires sci-libs/pytorch, the same dev-python/srt package mentioned above, and the openai-whisper Python module which also does not appear to be available in any overlay. It's homepage is https://github.com/openai/whisper So two new ebuilds will need to be written, but at least Python modules are easy. I may be able to do these if I have the time.
@bug wranglers: This is basically a new packages request. kde proj will not take on maintenance of foreign packages. @bug reporter: You should open separate new package requests per-package.
- dev-python/srt appears to be in guru overlay. - sci-libs/pytorch appears to be in ::gentoo. If you want to work on it, you'll just have to decide which of (VOSK|whisper) is easier/more futureproof to do.
