Matt Mireles, Great to be here 🙂. It would be awesome if TalkTastic could do real-time voice transcription while letting users keep talking until they're done. That'd be a great feature!
We actually had this feature in an early version of TalkTastic, but it was so frequently incorrect and inaccurate that we found it really distracting for users. Now that we have more advanced settings, it's something we could revisit. 🤔 I'm curious though - why do you want real-time transcription? And how would you feel if the accuracy was much lower than our current approach? We currently do contextual auto-correct or rewriting depending on the model's confidence level. I've been struggling with how to think about this feature. Would love to hear your thoughts!
Sometimes I have family calls with my parents and cousins in a big group, and I might step away to grab a parcel or something. When I come back, I'd love to see the conversation I missed. It'd be great if TalkTastic had that real-time transcription so I can catch up quickly. I'd just check TalkTastic to see what I actually missed while I was away. I wouldn't stress too much about the accuracy. I get that it improves after the question is finished(when I close the mic), but that's okay with me. And, it'd be helpful to have the option to turn this feature on or off in the advanced settings.
Lalit Kumar I use Granola for the exact use case you mentioned, but I sort of think it's distinct from why TalkTastic is so useful and I would want to keep those features in separate products
Fascinating.
Alex Libre Granola has mandated calendar permissions to work, which I find quite ridiculous.