Google Services - App for Translate and new Text-to-Speech for any device with url input

The flow below is an example flow I just made. Here you can see which cards I am using.

I select the URL in Cast card by click on the label en type: URL.


Remove the cast url from this flow.
Create a extra flow with the trigger from the tts device - Url triggered.
Then add the cast url action card to that flow with the local url token (first token in list).
Now it wil be triggered with the newest url.
Yours wil indeed sometimes be old because of asynchronised behavior (which would be correct, but not wat you want for this)

Ok I did and made the split.
I made an example of 2 different flows (Flow 1 & flow 2) and one output flow (Flow 3)
Can I use one flow to cast all the different flows I made in Homey?

Flow 1:

Flow 2:

Flow 3:

Hey Rob,

Not sure i am understanding you correctly, so let me first ask:

  1. Did you get it working?
  2. If you start Flow 1, does flow 3 gets executed?

Because thats how its supose to work, you setup a trigger (one time) that send the recieved URL from that trigger to a device: Your Flow 3.

Then al you need is a flow like flow 1 or flow 2 and start it.

So to me it looks like your ready to go and you should allready hear the voice when you start Flow 2 or 3?

Hey Arie,

It looks like everything works fine now. I have deleted all the cast card in the flows.
Yes, When I start the first demo flow, flow 3 gets executed.

I am now currently updating all my flows with contain the Google TTS.


And have a look at Universal Media Player - A Generic storages place for your Homey, speakers, tv’s or other media devices - Apps - Homey Community Forum (

Is uses the same mecanics (Start condition/action and then a Url trigger) to play sounds from a UPnP (sometimes “called” dnla) device (most/alott-of routers/wifi have the ability for that) on devices.

Use it as a condition card (and set the delay correct in the devices (also in TTS device=check settings)) and it waits for the sound to finish playing before it sends the speech.

That way you get a real got announcement system ( How to make a Great Public Announcement (PA) System with any device! - Apps - Homey Community Forum (

I am working on an App that wil make device-states (turning them on-off and reset-variables like volume) easier for the announcement system.

1 Like

I have looked to the Universal Media Player but unfortunately I cannot make use of this app. The Experiabox I use as modem/router does not allow me to use the USB port.

I do not have Sonos devices in my house but HomePods in all the rooms.There is currently no solution yet to use TTS for Apple devices (AirPlay 2). I still try to find a way to stream/cast (via AirPlay 2) audio to my HomePods. When I find a way how to, there is no need for me to use Google Nest Mini anymore.

This would be great! :+1:t3:

Is there any App for your HomePods?
Or not at all?

As fas as I know there is no app for it yet.

This seems very interesting, I’m currently using the “Google TTS” app to cast to my google mini’s, it’s very user friendly to setup, but it crashes a lot and I’m hoping that this will be more stable, but it is some more work to figure out :slight_smile:

I’m stuck at the cloud API account creation as it seems to only proceed with a valid creditcard. I’m from holland so usually no need for a creditcard as everything is IDeal around here and for international stuff I can usually use paypall, but Google is limited to creditcards only it seems.

I’m going to get one just to try your app :slight_smile:

Haha, yeah, its indeed a shame you need to fill in a creditcard.
But ideal is prepaid, and the point of the creditcard is that when you exceed quota’s you have to pay (and with CC thats is much easier for them then using prepaid or un-garanteed “checks”)

But i get an invoice each month per mail, and is still always €0,-.
So as long as thats keeps this way im happy. The current quota’s are way below normal usages, you could read out a couple of books each month (everage books is about 60.000 words averaging 7 letters a word that would be 420.000 characters (without SSML).

So lets hope they keep thesse limits for long :wink:

Any idea why this app won’t install (v4.2, 2016 model)? The old tts instals perfect

Yeah, its >=5.0.0., SDK 3, strange you can choose it in the Store. Should not be so i guess but. ← sorry, wrong

Okay thanks, will update this weekend and give it a try then

Your right, my mistake!
This is indeed (my only) SDK2 app, so actually i dont know why i doesn’t install.
But im still pretty sure that if you update your Homey, it will be fixed/work.
It did for other people. If not let me know please, then we can fix it.

On Homey 5 now, crossed my fingers and all Zigbee just kept working :slight_smile:

But now, done al the necessary things but can’t find the JSON download function… stuck now. Can you show me the way?

Nevermind, worked it out, not really a beginner thing to do. Several different steps must be done in the right order. Could use a bit better instruction for noobs. Having to experiment because homey doesn’t support voice commands anymore, trying to let my new Google Nest Mini do all the feedback when events happen, talking, including all the messages and reminders for the kids etc.

Can you specify the pro’s and cons for the two card options Google Cloud TTS device VS Google TTS Device (non cloud)?

Ofcourse cloud, non cloud but wich one is faster? Takes less space? Need connection or not etc?

Hey RWin,

Indeed, setting up this app unfortunately is not very easy no. I still wanted to share this with the community, because at least (developer-)users who really want this can have it without having to program it all.

There is a big difference between Non-Cloud and Cloud

Non-Cloud is retrieved everytime from the (old) google translate/tts server, with only a language code (one voice per landcode, non-chooseable) with speed mode normal or slow.
Also it doesn’t support SSML.
I build this for people who dont want to get Google Cloud, but do want a simple TTS send to devices that take URLs, like TV’s, for which the current Apps (ofc) don’t provide TTS.

Cloud is build way smarter, having a memory of the last 100 transactions (putting every request (back) to the start of the memory, so the ones you repeat are “saved” in memory).
You can pick multiple voice per languagecode/country, you can adjust pitch and speed (in percentage up and down, not just normaal/slow). So Cloud will use less connection, but to be honest, both speech deliver very small files, mostly just a few KB, i mostly build it because it saves a lott of monthly characters (allthough i don’t even get close to the limit) and ofc also for the speed.
Also supports SSML which makes using this in Homey way easier with a much more natural output.

If you like me want a real P.A. system through the whole house with a good accepteable voice, Cloud is the way to go. :slight_smile:

Okay, thanks! I completely killed Homey speech now :slight_smile: Everything runs through Google Mini now. With the Google routine function i can let it trigger a flow with specific words. All messages homey gave are running through your app now.

Would be great if Athom adopt this function and users don’t have to go through all this work. Will cost Athom only a few bucks for the API usage.

Two other questions, is there a way to remove the notification before each message? Also the volume is around 25% less than the level set within the Google Mini / cast device. Can you solve this somehow?