Nerd-Dictation - a simple, hackable speech to text tool for the Linux desktop
Posted by ideasman_42@reddit | linux | View on Reddit | 3 comments
I had never been satisfied with any of the dictation tools available on Linux, until recently where I found an open-source speech to text engine and gives excellent results, however it is just a library (VOSK-SDK).
So I put together a small script that integrates it and makes it a tool that can be used for dictation on the Linux desktop, I use this with a bare bones tiling window manager only activating it when I want to do dictation, so no background processes.
While I realize this probably isn't enough for everyone, for basic dictation (including this post) I find it sufficient.
Professional-Crab234@reddit
There is a new kid on the block:
VOXD - a voice-typing / dictation app for linux
"Out of the box" sets you up with LOCAL voice transcription, and even LOCAL ai-rewriting according to your custom pre-made prompts.
Works on CPU. No GPU required.
Sudden-Lion9886@reddit
Is it possible for nerd-dictation to ignore speaker audio and only listen to microphone... the problem right now is that if you play music or are hearing to a video, that audio is double captured via the microphone
ideasman_42@reddit (OP)
Not via nerd-dictation, it could be that starting nerd-dictation pauses/disables other outputs, re-enabling sets them back to the previous state. But this is something you would have to configure.