wow, impressed!

[atomGit]

last time i used any STT program was a very long time ago and you had to train it and it didn't work well at all regarding accuracy

i was expecting to have to train Transcribro as well and despite there was no such step in setup, i was very pleased with its accuracy - how exactly it's able to recognize my speech so well, i don't know, but it seems to work great

the delay in output is slightly bothersome, but certainly acceptable given it does its job far faster than i can type

anyway, just wanted to say thanks to the developer(s) and contributors! very nice :)

[soupslurpr]

Thanks! Whisper (the speech-to-text model used in Transcribro) is trained on audio and text pairs scraped from the internet. This is why it's very robust to many voices.