4 comments

[ 0.25 ms ] story [ 26.7 ms ] thread
I set the price to free but it will take some time to update. Comment from a voice memo I did just now:

   I made this because I do a lot of voice memos.
   I love to just riff, you know, just speak my mind and say whatever I think.
   And not kind of worry about, oh, I need to write this down.
   I need to remember this.
   I just love to kind of say it in the moment.
   And I come up with a lot of good ideas like that.
   I get to clarify my thinking.
   And so I created this to just provide a record.
   You know, if, what did I exactly say there?
   What was the kind of flow of logical?
   What were the points I was making?
   What were the wording that I used?
   And then I can go back and I can check it.
   And I don't have to waste time listening to, you know, a voice memo, which may span over
   hours in some cases, you know, with a whole lot of silence in there.
   And I can just set this thing to transcribe it.
   And it makes it easy.
   So that's why I made it.
The quality is waaay better than the normal speech to text voice typing you get on Google keyboard or MacOS. Which is so relieving to get effortless almost perfecdt quality.
I should say it works waaay faster than real time. There's two modes: Turbo (around maybe 200 words per second, depends on content) and regular (maybe 60 words per second, eyeballin' it). It uses OpenAI's whisper (actually whisper.cpp) and Silero VAD (for the removal of silences / non-speech).
Love it. Just downloaded and used it. Would you consider adding diarization to your roadmap? I am not sure how hard it would be, as other implementations need to use hugging face/pyanote to do it, instead of local processing. (I left a review on the Appstore also).

I propose an in app purchase for diarization (and any other requested features).