With the limited vocabulary needed for automation control, why not use a voice recognition program in a PC, and do everything locally?
Your kidding right? I've been using an AHP add-on software
I found here at the forum way back in 2005 Tuicemen and others have AHP compatible voice control software. I thought we all were using some level of voice control already.
Although I really like my
BVC (Bill's Voice Commander) software.... it was never updated to overcome the problems that came with Visa. Which makes it a great software to run on an old XP dust collector or even an old CE mini. I purchased a 3rd party Voice (Kate 16bit) for my set-up. And (with a little speaker tuning) my automation announcements sound very human-like (much more so than say... the Hilary voice).
My AHP/BVC set-up announces everything from a (waking hour) clock announcement, outside perimeter detection, garage door status.... and more.
I can voice command my X10 stuff on and off too. I had come up with the idea of using baby monitors instead of an expensive wired microphone set-ups. And partly because it's pretty quite around here with just the wife and I. The 8bit voice recognition works great with simple $20 baby monitors.
But there are serious limitations with local control! All this has been covered up/down/and sideways here on the forums
for over a decade.
I have two laptop automation computers running 24/7. One is totally local and has the power to do a GREAT job with voice recon. The other is simply a server to talk with a automation app on my phone (and a CM19A). I have it all setup together with my (security camera) video recorder/server on a shelf in my Man-Cave (formerly called Nerd-nasium, or home office).
Don't get me wrong...
it's a nice set-up. But technologically speaking.... it's a
blast-from-the-past too. My cell phone could easily accomplish all the same functions. And without the concern for power spikes or electric use. Heck.... for that matter.. this could all be done on a $50 tablet. AND with a CM19A (type device), a tiny server app, a phone app, and a cloud portal... I could have the-whole-nine-yards with TWO-WAY global access.
And... NOTHING compares to the cloud based voice recognition stuff. I use Siri to type my text messages.... and even to spell words for me. I can set in my easy chair at home... with my iPhone in it's charging cradle (which give me a great view of the screen). And say: Hey Siri, text Mary (wife) and remind her "
tonight is pizza night". Siri will repeat the message and ask confirmation to send.... and it's done. Without so much as lifting a finger. (I could easily order the pizza the same way as well).
Using the API cloud function like
dhouston suggested the other day is the perfect (complete) automation solution. But it requires a cloud connected X10 interface. But like I posted earlier.... the device doesn't need be anymore than a cheap tablet, a CM19A-like dongle, a couple apps, and a cloud/server site.