<![CDATA[Yet Another Voice Module -- I want my MM to speak now!]]>I’m very frustrated by all the different and poorly maintained/documented MM voice modules “out there”. So to make matter even worse, I’m proposing to create yet another one!

However, this will be “The One” that does all of the above below! Meaning it will work, and be easy to install and well maintained.

What would the features of this MMM be?

  • It would concurrently accept wake words for: Google Assistant, Alexa and Snowboy (for MM control).
  • It would be more modular so that other new speech technologies can be implemented later, (such as Bixby, Siri and others.)
  • It would focus on low CPU usage and memory foot print to be ideal for embedded RPi3 (and above) devices.
  • It would be properly integrated so that it is also able to respond to future requests, like:
    • Wake me up tomorrow at 8 am.
    • Set the timer for 20 minutes.
    • …and various follow-up questions, to previous responses.

The project would probably not need to be built from scratch since there are a very few modules that more or less work (such as MMM-Assistant) but they need a lot more TLC, so we might need to fork it up to a higher level of maintenance and perhaps try to repackage it into something more fresh.

What do you guys and girls think?
Who’s interested?

]]>https://forum.magicmirror.builders/topic/6810/yet-another-voice-module-i-want-my-mm-to-speak-nowRSS for NodeThu, 18 Jun 2026 11:46:20 GMTTue, 20 Mar 2018 12:11:35 GMT60<![CDATA[Reply to Yet Another Voice Module -- I want my MM to speak now! on Wed, 21 Mar 2018 13:23:57 GMT]]>I Really like The functionality that comes with Alexa … However I Really hate To give out My Data to a company
So My Current Attempt Is a offline Voice recognition Called snips that Works pretty well for me (I use a 4mic Array And it can recognise what I say from a distance of 5m with background Music without any problems) in Basic in snips You can create own assistants pretty easily (for me one of The Main reasons to use it Is that The Automated Speech recognition Is Not only avaible in English but for example in German Aswell)
after recognising a intents it sends it via mqtt which You can acces with Most IOT devices or for MM2 there Is already a MQTT module

Well Thats My current Attempt
While I Really like that it I can “easily” use it for everything in My house And have German As language The big disadvantage Is obviously that it Is a Second software And so definetly Not Really a one Click installation Thats Really easy To use for every User that just wants The Thing To work without any hazzle

]]>https://forum.magicmirror.builders/post/37272https://forum.magicmirror.builders/post/37272Wed, 21 Mar 2018 13:23:57 GMT