Use Google Home Speakers as Smartthings Audio Notification Speakers

Ignore what was written above, it’s not a Google Home limitation, it’s a SmartThings limitation, they just don’t want to do it, you are correct.

1 Like

I have cast-web-api setup on my Raspberry Pi 4B to use my Nest Mini for certain ST automation announcements but the functionality is limited - can’t control announcement volume (always 50%) and it won’t resume if something was already playing on the speaker.
You can find it here [RELEASE] cast-web v1.2.1 - Chromecast Device Handler & SmartApp
Also, there’s an installation video tutorial here How to Control Google Home in SmartThings | Making Google Speak in 2021 - YouTube

4 Likes

I am using Samsung R1 speakers, but nowhere as good as when Alexa Speaks worked. I am hoping someone comes up with a way to use the Echo devices again.

You can use echo devices to speak whatever you want by having an action in ST trigger a routine in Alexa. I have mine speaking when doors open or outside motion occurs. Not as robust as Alexa Speaks but do 95% for my use cases.

Several threads here show how to make this work and all you need to do is search the forums here.

I created an Alexa virtual switch which shows in Alexa. Then I created a automation that is like this.
If:
Hallway Motion detects motion
if:
Front door contact is opened
Then:
Close Front Door Exit (which is the Virtual Alexa switch.

I then went into Alexa app and wanted to create a routine using the Alexa virtual switch, but it isn’t available in routines, but I can see it in Alexa devices. More info on what I am doing wrong would be helpful.

It can’t be just a virtual switch because at the present time Alexa does not allow switches to be the triggers in routines.

You have to use special code for a virtual device which is BOTH A Switch and a Sensor. When you turn on the switch, the sensor will look like it opened. when you turn off the switch the sensor will look like it closed.

Since Alexa can use sensors as the trigger for routines, it will work.

This is all explained in the community FAQ, including links to get that particular code, so let’s not take this thread any further off-topic, because this method does NOT work with Google devices at this time.

FAQ: Can I trigger an Echo Action without Speaking to It?

I might have missed it, but I can’t find the code you are referring to a virtual device which is a switch and sensor.

There’s a heading in the FAQ:

See the following How To article in the community-created wiki for details, including a DTH if you want to use the combined sensor/switch method.

Follow that link and you will find a link to the code.

OK, I found it but the delay from opening the door to hearing is long, at least 20 seconds. Also, I am on a 1gig Internet.

This painfully slow, I really miss Alexa Speaks.

Smartthings is slow when it comes to integrating new devices. I m an electrician and my customers dont get my recommendations.

you can get the ikea symfonisk speakers for $99 from ikea or looks on ebay to get them cheaper like i did. they work as sono speaker w/o AI.

It’s not easy but I do that using NodeRed, a software installed on a Raspberry Pi (you can install it on any computer provided it stays on).
NodeRed integrates with Smartthings through a palette (on NodeRed side) and a smartapp (on ST side).

Node red has a palette for Google Home to cast TextToSpeech on Google Home or Chromecast devices.

So you can either.reat flows in nodered triggered by ST devices, or you can create automations in Webcore that trigger NodeRed Flows (e.g. a TTS action) through a Web POST that has inside as variable the text you want to pronounce.

Do you have detailed info on the setup for using this in ST?

I did a tutorial, it’s in Italian but you can probably follow the screens:

The NodeRed palettes are linked in the description.

The guide to setup the Smartthings palette is in the documentation of the palette, if you can’t follow the tutorial. Be careful that when in the smartapp you input the nodered machine IP (the raspberry pi in the tutorial) you need to add 'http:'before it.

In the tutorial I show the usage for the Alexa palette, but the one for Google is even easier, you just need to configure it with the Google Gome’ s IP address, and provide it with a payload with the text you want it to speak (it will make sense after you installed NodeRed).

Let me know if you’re stuck somewhere and I’ try to help.

I tried to watch, but I was lost within minutes. Thanks for the offer to help, but I was lost at the beginning.

It’s really pretty incredible that there is no inexpensive, easy-to-integrate solution for audio notification speakers. Many of us have had to cobble something together ourselves. Mine happens to be a Raspberry Pi connected to some powered speakers which is now implemented as an Edge device. I’m actually contemplating putting together some kind of package that I could sell to the community to solve this problem. The trick is finding a decent-quality single, powered speaker that could be paired with a tiny and cheap Raspberry Pi Zero. This would give you a wireless speaker device running locally with no lag. For TTS, right now I use IBM’s Watson for it’s super high quality and and I cache the audio files locally so it only has to go to the cloud when you request a new phrase that hasn’t already been generated before. So most of the time it all runs 100% local with instant response thanks to the new Edge integration. My vision would be to have a $30 small, integrated box you just plug into an outlet wherever you need it and it would be discovered and fully integrated with SmartThings. Certainly not high fidelity for music, but good enough for audio notifications.

I’m just not sure how big of an interest there would be for something like this. For anyone curious, I got one of these tiny speaker modules for RPI from Adafruit and will be experimenting with it to see if they would be powerful enough for a solution like this.

5 Likes

The solution WAS the Galaxy home mini, which seems to be cancelled, but I still have five of them and they work fine for this exact use case (plus integration of certain IR devices)

If you can find one, just buy that /shrug

It’s as simple as making an automation and selecting the home mini, then you can enter your announcement in English or Korean.

My Echo devices respond pretty quick to Alexa Routines…Almost instant. I was using LANouncer which was pretty darn good when paired with a tablet connected to dumb speaker. Actually, that worked really really good. Then came EchoSpeaks which was absolutely AMAZING!!!

Since device count in ST is limited, I had to delete alot of my (non essential) Virtual Devices which cut out alot of the chatter on the various speakers throughout the house. TBH, Alexa was speaking way too much around here and it annoyed more people than it helped. :rofl:

Following… need a tester? ST hub v2

Hey there @7522. I was experimenting with the speakers from Adafruit I had mentioned. They work, but probably aren’t going to be loud enough. I also came to these realizations:

  1. Hitting a price point of $30 or thereabouts is almost impossible without using pure junk. By the time you add up the cost of a Pi Zero (prices have gone up!), the speaker kit, a power supply, and an enclosure, it gets you up around the price of other cheap powered speakers already available, albeit that wouldn’t have the TTS integration.

  2. I thought about getting a 3D printer to create a custom case, but… that ended up looking like a rabbit hole I didn’t want to go down

  3. There’s a configuration challenge where you need access to a good text-to-speech service. As I had mentioned, I personally use IBM Watson. Although it’s free for the amount I use it, expecting a user to have to sign up for an account and go through the configuration that would be necessary would severely limit the audience. I don’t think Amazon or any other TTS service would make it any easier; they all require sign up and configuration. I haven’t really looked to see if there is any open source software available to perform this function locally. I also thought about just including a bunch of canned announcements, and making the ability to get custom ones an option for the intrepid.

So given all of the above, I haven’t progressed on this project, so am not looking for testers right now. But thanks for expressing interest! If you have any other thoughts on this topic I love to hear them.