Speech input and recognition

[Massimo B.]

Is there anything usable on Linux to have a Speech Input option on any edit field where keyboard typing is possible? Also supporting for more languages beside English, like German?
I only found
app-accessibility/julius (https://github.com/julius-speech/julius)
https://github.com/mkiol/dsnote
And an entry at Wikipedia.

This option today is available on systems like Android and MS Windows, often using a cloud service in the back which might be another issue.

[undrwater]

It's be a little while since you posted, but hopefully you find this post somewhat helpful.

I've been on a similar quest, and there are a few projects based on the Whisper-ai project (LLM for speech-to-text), but no ebuilds as of yet. These run in a python venv, so that's how I'm trying them out:

https://github.com/dhruvyad/uttertype
https://github.com/savbell/whisper-writer

Any language that whisper supports should work for these projects.

Additionally, I'm using piper for text-to-speech. Also in a python venv, and I've got a keybinding that reads out highlighted text.

Good hunting!

[wildhorse]

speech to text
https://github.com/mozilla/DeepSpeech

text to speech
https://github.com/mozilla/TTS

text language translation
https://opennmt.net/

You could contact researchers at Fraunhofer IDMT and Fraunhofer IAIS as another source of information.