Play with audio commands

[Soloam]

Hello, I've been gathering some info regarding adding voice commands to my mysensors! My idea was to have a low level microphone node, a simple arduino that would stream the audio to a server running linux that in the end would capture that sound and process it (recognize it) and would notify my main controller to do some action.

I found this very cool app http://www.bitsophia.com/en-US/BitVoicer/Overview.aspx that it seems to do this, but there are several problems. One only runs on Windows, two, it's not open source project (that i would prefere).

First of all, any one know ant way to stream real time audio from a arduino to a server running linux? I found this http://tmrh20.github.io/RF24Audio/ but I still need to test it, any one used it?

The second part would be to recognize the sound and make it trigger some action. This still needs some digging, but nothing that good old coding would'd do! I just need the string with the commands in text.

Other alternative would be to use a Voice Recognition Module for the Arduino ( http://www.elechouse.com/elechouse/index.php?main_page=product_info&products_id=2151 ) I don't like this approach because I would be limited to the commands in that arduino, a streaming solution would bem more modular.

Any one had ever played with this?

Best Regards

[Dan S.]

I ordered an Echo Dot from Amazon. Won't be shipped till July, but hope to integrate it with my Vera controller to activate both my Vera z-wave and and Mysensor devices via voice control using the device.

[Soloam]

I'm studding the possibility of making the speech recognition with a raspberry pi zero and google stt api! Then I would publish a MQTT string to control my MySensors with OpenHab. The main problem of this approach is the delay time between the request and the action, because the audio must be translated by google and return to analises.