The question is not quite clear, but you will find everything which is available in .NET FCL, here:
http://msdn.microsoft.com/en-us/library/gg145021.aspx[
^].
I did enough of both recognition and synthesis using this assembly (System.Speech.dll), it all works. Recognition is not so good, but it quite reasonably good for small
grammar. For
dictation grammar (if you wanted to dictate arbitrary text with wide lexical repertoire), recognition is nearly a complete failure.
If you pronunciation is off, recognition is pretty poor. Will this fact be useful for your purpose? :-)
—SA