It's great and now I am sharing few scripts.

[gnusupport]

While I am typing this in reality, I am speaking and it is getting transcribed straight into the browser window. It's the script I made today which accepts the speech and then it gives it on the standard output. Hope that this script is going to be useful for people who wish to quickly create the standard output transcript from their speech.
https://gitea.com/gnusupport/LLM-Helpers/src/branch/main/bin/rcd-llm-speech-to-stdout.sh

Often, I am using this type of script, which is then basically invoking the speech listening, then it transcribes it and provides me with the typing by using xdotool:
https://gitea.com/gnusupport/LLM-Helpers/src/branch/main/bin/rcd-llm-speech-single-input.sh

Hi, I was using Nvidia Canary 1 billion model, but since I found that whisper.cpp is there, I'm using now the combination of those two. If the Canary is running in the memory, fine, but if not, I am running the whisper, which is really free software.

Great to speak to computer and get words typed because this is not common as we don't have on GNU/Linux such a good integration like on Android phones, but now it works so well it works so fast I am very happy for this.

Thanks much to author who made it simple to run the whisper on my computer.