Voice Control for Smartthings


I just received my Smartthings starter kit and I am in the process of making the transition from my Zwave stick + computer to the Smartthings hub.

In the past, I was using a Zwave server and other people had used SiriProxy or other solutions to be able to control their home automation network with voice commands. I’m new to the Smartthings community - but is there anything that would allow for voice enabled controls of the Smartthings network?


1 Like

they are building ST support

Thanks Coolcatiger. I’ve seen Ivee along with other kickstarter projects like Ubi and Fiona (failed to receive funding). But I don’t see that these products are available right now.

I’m looking for a solution that is available to implement immediately. Perhaps Siri proxy?

Tasker support for SmartThings would be a very “nice to have”; as that might permit interfacing through Android/Google voice input.

1 Like

ivee uses this it looks like: http://developer.att.com/developer/forward.jsp?passedItemId=12500023

Yeah, Ivee looks interesting. Fiona looked pretty ridiculous, but I guess it doesn’t matter since it wasn’t funded.

Ubi isn’t available immediately, but is shipping beta units now and should be shipping production units in a couple of months. I think November or December is very reasonable given their progress.

They will actually have working SmartThings support, though:

But to answer your question, I don’t believe there’s anything you could use for voice control of ST right now.

I mentioned this in a different thread, but just to repeat it: I’ve been a bit reluctant to back voice based interfaces. I’m just not sure the technology is ready yet. Now, I hope I’m wrong and am able to get something later that will do this.

But I look at what I have on my phone right now: It works… mostly. I can give voice commands to my phone and it, more often than not, get it right. Yet it’s still far from 100% accurate. Furthermore, I find it needs ideal conditions. If a radio is playing or if there’s a lot of road noise this interferes with the voice commands. If kids are talking or my wife decides to ask me a question half way through a command (happens more often than you’d imagine!), then it doesn’t work.

Then I think of my home environment. The Ubi or Ivee likely isn’t going to be three feet from me usually. It will usually be 10 feet or more. Can the device clearly hear me at that distance? Then the TV is often on. And someone might be on a phone call. The dog might be barking as someone walking by the house or kids are fighting about who gets the last cupcake in their lunch. Maybe the ceiling fan is running which adds a touch of noise or football camp is going on in the park by my house and the coaches are constantly whistling.

Long story short… there will rarely be ideal conditions for issuing voice commands. There will almost always be some background noise and there will almost always been some distance between me and the receiving unit. I just don’t have confidence yet that the technology will be able to filter out other voice and noise and listen just to what I have to say.

I mostly agree with you Chris. However, A plug in devices is going to have a lot of advantages that a smartphone doesn’t. It can boost a lot more power to microphones, and could potentially have a lot more local processing power (however the Ubi doesn’t). Considering how well this year’s phones are doing (especially the Moto X) I wouldn’t be surprised at all if we saw something even more dramatic if it had an unlimited power source.

Well, as I mentioned… I hope I’m wrong. I think it would be really cool to be all “Star Trek” and say: “Computer, turn off the lights.”

I just don’t think we’re there… yet.

I use Vera currently, and the app I use to control my system is AutHomation. This app has voice control functionality baked in, but you have to open the app and tap the Voice option. SmartThings needs to integrate this into their app as well, then improve it by integrating with Pressy for quick “press-to-talk” voice control so you don’t even have to look at, or touch, your phone screen.

In addition, they should develop a bracelet or necklace that would connect to your phone (specifically the SmartThings app) via bluetooth for voice control without even having to hold your phone. It could be like a little Jawbone device but connects to the SmartThings app directly.

Also, by using a bracelet or necklace you can bring it to your mouth while giving a command, which would solve a lot of the problems you guys just brought up.

@cnguitarist (Brandon):

Very good observations.

There are expensive “custom” (through authorized dealers only) home automation systems that are “always listening and built-in”; these require quite a level of sophistication to be able to avoid false triggers …

I just purchased a Moto X phone which uses “always listening” technology; but it is tuned only to one specific person’s voice. That, along with reasonable proximity, seems to avoid false triggering, though I have not tested extensively.

Somehow, I have trouble believing that avoiding false-triggers is an insurmountable challenge, particularly if we are willing to accept unusual trigger phrases (e.g., instead of “OK Google Now”, or “Hey Computer”, why not “Open Sesame Do This” … – in practise, linguistic databases could be extensively searched for sound patterns that are unlikely to occur from television broadcasts or normal conversation or noise).

I am a Beta Tester of The Ubi (http://www.theubi.com) and my primary issue at the moment is false triggers; but I also am aware that this is a high-priority focus of the developers. Then again, I wish they could afford to license Google Now and/or Moto X’s various sound processing and voice phrase handling technologies.

And… then partner with major HA systems, such as SmartThings.

Lots of opportunity here for those willing to invest in the challenge. At the moment, this research is costly and results in expensive end-products (if any).

Tasker for android can fairly easily be setup to control smartthings using endpoints.

Autovoice (a voice recognition/launcher plugin for Tasker) has just been integrated into Google voice search by an Xposed module.

It is especially cool with a moto x phone as you have access to the always listening by saying “Okay Google Now”.

See video for demonstration (I beleive he is using Veralite/Autohomation but Smartthings/Endpoints would work the same):