> Um, because, all other things being equal, you'd have to parse out an audio > stream to get at the information you'd need...? Having just sent myself a voicemail and then listened to it, it seems that the messages aren't denoted by tones like I had hoped. Knowing what to record, then, would be rather difficult.