It's been a long time since I've used Sphere and I'm curious if it can leverage any form of speech recognition or speech-to-text.
The short answer: no.
The longer answer:
Regular Sphere doesn't, since it doesn't have any sort of sound recording.
TurboSphere probably could, if someone actually decided to write a (C++) plugin that accesses a speech-to-text library. But TurboSphere is still in beta. It also would require extra effort to write that plugin, and it would probably require some more effort if you want to make it cross platform (different STT engines?).
Same as DaVince said, short answer no; but it's not impossible for it to happen. TurboSphere's audio library of choice, BASS, has good capabilities for recording. But they aren't being used as it is.
So if you want to play with C or C++, learn a little bit about V8, and find another library to turn the recorded sound into text, then yes. But it would not be easy.
The TurboSphere plugin interface is cross platform, especially if you use the tools in 'plugins.h'. You could write a platform-specific plugin, but there's nothing built in that causes that at all.
Well, I was thinking it would mainly be the STT engine being the determining factor on how portable it is. I really have no clue if there are any out there, which is why I said that. Though maybe Google's has an API for that (though you'd have to send the speech data online if you do that)?
Kinda also begs the question "Which speech recognition is better: MS, Apple, Google, Nuance (Dragon Naturally Speaking), or some random 3rd-party's?"
I am by no means a C or C++ developer, but this sort of thing is interesting so I might have a look into it. But not sure on how portable it would be. This is just experimenting really.