I don't know if it's my audio setup or my voice, but even with a lot of effort training Windows speech recognition, it generally takes VA 5-6 repetitions to react to even a very short spoken command; anything longer than that is hopeless. I use keyboard shortcuts instead, but there's a very limited space of key combinations I can use without interfering with other applications.
Ideally, what I'd use is a sort of console interface - one keyboard shortcut to bring up a text prompt, and whatever I enter is treated like a spoken command.
Is there any plugin that can do that?