like say have speech recognition put what I say into the google textbox

If you want that then you'll need something designed for that purpose rather than something designed to control a robot. Windows Speech Recognition can do that but it's not too great unless you have an amazing mic and have trained it a lot (and I mean a lot). Dragon is the best of the bunch for that.

You could probably do it through Pandora Bot, EZ-Script and a lot of work. Pandora Bot is the only part of ARC which supports dictation, all others require a phrase list.

yea I was looking into the open program command

Exec("c:/.............")

then have it sleep for 30 seconds or so and then terminate the speech-to-text program

What programming would you have it open though? And how would you get it to take the payload and use it in the right place?

It all seems like going about things the long and messy way when you could achieve much better results through other software. You can pass information to ARC from other software quite easily using telnet, a lot of the fun scripts and actions I added to Melvin tonight use other software and pass results and variables (which act as triggers) to ARC through telnet via EventGhost's python scripting.

unfortunately I don't quite understand python scripting. I've read all the documentation but I just can't seem to understand.

I'll let you in on a little secret... neither do I... but it's easy enough to adjust an example to send the right info through telnet