Hubitat with Snips and Matrix Voice

Had anyone tried to integrate these 3 platform with success?

It looks like the ideal project with everything offline!

Check:
https://www.matrix.one/



https://snips.ai/
Wow!!

1 Like

Good idea; however, may be overcome by events with Snips purchase by Sonos. It would (of course) be a pi or similar device with mic and speakers.

1 Like

Exactly but we can be controlling our hub devices with offline voice (the main problem I see nowadays with google home).

Matrix not only can control using voice but gestures as well as it have a camera kit included and it can trigger events based on hands movements, face recognition, etc.... it is very interesting to control our HE with all those . :stuck_out_tongue_winking_eye:

1 Like

But the snips website now just links to the sonos press release. The website is basically shut down. So, I don't know how you would do this anymore. Also, the release you linked to says that it would have to run on something stronger than a Raspberry Pi, so probably a desktop or laptop. To me, the amount of work necessary to get this up and running sounds like to would be way too much work give the readily available Alexa and Google Assistant.

Have you seen this thread? Requires a Windows PC, since vox commando was originally designed to work with windows media center.

2 Likes

Why hasnt anyone managed to hack the alexa/googles so we can direct them to a local server instead of the google servers? Maybe get the mic and speakers controllable over wifi, then we could use the input and output for literally anything we wanted... Google Documentation is pretty extensive, and already allows a lot to be done, and these things are cheap enough its not like destroying a few will hurt anyone financially so why isnt this a thing?
Is there a project anyone knows of that is already working on this (specifically the googles) ?

2 Likes

You're right but there is a workaround to this I think... check this:

and this:


I just started to read about this but maybe Rhasspy Offline Voice Assistant is a good alternative and work with Matrix voice as well.

I just posted here when I saw about these projects... I have to make a whole research to be sure what way to follow but one thing is sure for me:
I need an assistant who do not depends on the cloud to control my local devices, it was the main feature I loved from Hubitat and a good assistant I think is the one who works locally and with the cloud, as HE do

Any ideas will be welcome ( marktheknife I will read your suggestion for sure in a few minutes)

I have a dream (with apologies to Martin Luther King).

  • A small (echo-dot size) device (aka HubitatPuck):
    • Mic with far-field voice reception and processing,
    • embedded (local) voice processing to convert input speech into device, commands, and parameters,
    • for simplicity, does NOT know the state of Hubitat devices,
    • User identification of devices, commands, and command parameters via a JSON file upload and web broswer.,
    • A MQTT or other "broadcast" interface that sends commands to connected Smart Home Devices.
  • a Hubitat Application that
    • Allows selection of devices and capabilities to link to the HubitatPuck,
    • Receives the json formated command fron the HubitatPuck and executes on the devices,
    • Returns success / failure for executed commands.

Oooopppsss. Just woke up to reality.

Dave
"Set my peeps' voices free from the cloud."

4 Likes

HubPuck :slight_smile:

1 Like

good, but I was thinking of expanding to SmartThings and think SmartThings having a Hubitat application would be funny.

2 Likes

Hi markthenife.. I just read your article suggestion and I have the Voxcommando making test.
So far looks good...

I would like to use some Wifi Mic like Matrix Voice Mic Connected to the same server as they do with the Pi... I'll continue researching... Thanks for this man!

2 Likes

I’ve completed a basic Python app that listens for specific Rhasspy voice commands (called "intents") via its MQTT event service and based on the voice command sends commands to Hubitat via the Maker API app.

Once you have Rhasspy up and running, the following are the main steps:

  1. Write the Rhasspy configuration for sentences for voice commands.
  2. Select the devices in Maker API you want to expose for Rhasspy control.
  3. Add the intent rules/code in Python to control devices in Hubitat.
  4. Set up the Python code to run as a service on the Raspberry Pi (or whatever you are using to host Rhasspy).

I have examples in the Python code to get started. I’m using the pyhubitat and rhasspy-hermes-app libraries to simplify the integration.

The Github repo is here.