Cool, I once used the MS Speech API. For text to speech and speech to text.

Let us know when you have a working example .