What voice related modules are available?
There seem to be many different voice modules available, but often looking at their repositories they have not been updated or maintained for a long time.
I’m looking for a working module to:
- Have an Alexa-like functionality. (I.e. Asking a question and getting an answer.)
- Having a MM voice control, to switch modules or pages
What are you using and how well does it work?
Since, you know I love lists… perhaps we should make a Wiki list of the different voice modules, their development status, quality, dependencies etc.
| Name | DevStatus | CustomWord? | UsesAI? | PrivateData? | ConfigDifficulty | MM-Quality | Description | | Lola | fake | yes | yes | no | easy | great! | An uber-duper SHE OS module that solves the question of life |
give or take…
Here are some:
There is Kalliope assistant too..
- self hosted TTS and STT
- framework to build your own assistant
- multi languages
- easy to install (rpi image available)
- a lot of plugins (neurons)
- android app
Here is a demo of some modules:
And the demo of MM integration:
Have fun !
Chris last edited by
Concerning the voice modules, I tested the modules MMM-Google-Assistant and MMM-Assistant for the “interactive” part on the internet, I never managed to make them work. I asked several questions, but apparently I must not be the only one because I did not have many answers.
For the “non-interactive” part, I use the Jarvis solution (https://www.openjarvis.com/) coupled to the MMM-Jarvis-Voice-Control module. the result is correct and there are several interesting plugins.
Chris last edited by
I went to see the Kalliope site. It sounds interesting.
Regarding plug-ins working with API Google, can you tell me if there is only the translation function available (plugin translate) or if there is a plug in approaching the principle of “google home”?
Thank you in advance,
Please help fill in the details for your favorite voice modules!
Hope you guys like. So please let me know if there is something you don’t like or that can be improved.
@E3V3A thx !
Nice post idea. This should help a lot of people.
I started my mirror project by setting up google assistant first by following this guide:
The guide is fantastic, there was not one hiccup with the install, and the GA works as expected. Currently, I have it running with in conjunction with MM without integration. The integration would be nice though.
@StacheEnthusiast That is a nice instructable, but mostly useless for us MM people. The reason is that it relies on a pre-compiled image:
Next we need to download the Voice Kit microSD card image for the Raspberry Pi.
So there is really very little to learn and will surely not integrate with MM, as you already stated.
The biggest hurdle is always the sound config and the super non-intuitive bureaucratic Google registration process, requiring you to give both phone numbers and credit card numbers, and countless of impossible options while not being able to navigate through their pages in any sensible manner. (Eg. You click a link, then you realize it’s the wrong page, and you hit back button or the back icon, and WTF, you end up at a different page than you came from! Googleism!)
At the end of the
dayyear, the biggest hurdle of any of the currently available voice products I have looked at, are:
The useless (at best) or complete lack of installation instructions!
I have no idea what is going through the heads of all these developers who spend weeks and months of hard effort to produce software, that nobody will use, just because, anyone else who are not exact Matrix clones of them, will not be able to install, nor use it. Please understand that these are all wasted efforts if you’re not interested in having other people being able to use it.
I’ve been around long enough to confidently say that the vast majority (~98%?) of failed projects on github are only because of lacking usage instructions and lacking participation of the original developer or repo owner. I’ve seen and follow many great projects and ideas out there, but most will die a slow death of ignorance in the invisible shadows of murky README documentations. People come, look, move-on and forget.
In the current context of voice interaction to your MM, this is particularly disturbing. Although, many MMM’s here are fairly well presented in this forum or in their READMEs, the voice related ones are extremely lacking. This was a bit surprising to me, since people were able to voice interact with their C64’s already in 1985. While 99% of you probably have no idea what a C64 is, its not important. What is important, is that one of the main differences was that SW in the old era was usually extremely well documented. It was a sales advantage. Even yo’ mama could understand it!
So if you have spent countless hours thinking about and developing this great module that could improve the life of all MM owners, and you have a this l33t group of really super awesome developers and hackers, that project will never really make any significant difference, unless you can make it accessible also to the novice, whether or not this person is technically interested or just a plug-and-play grandma.
Here are the most important points in writing documentation:
- Don’t assume any previous knowledge!
This is by far, the number one doc issue and failure!
- Does the software depend on any other, previously installed software?
If so make sure to either re-iterate the steps or link to equivalently good installation instructions.
Specify exactly what it depend on, and where to find it.
- Are the steps following a logical order of operations?
You’ll be surprised!
- Is the information you give, correct?
Don’t guess how things work!
- Are all the steps up-to-date with current master repository code and functionality?
This is the hardest to keep up, since most SW in github are under some development.
- Are your screenshots up-to-date with how the UI actually look?
You have screenshots, right!? A million of your words, could never replace them.
- Have you started from scratch and followed your own installation instructions?
This step is what fails 99.9% of all documentation. Unfortunately developers are incredibly lazy in this regard, since they think it’s a waste of time. I’m sure NASA has something to teach you about this.
- Can anyone complete the steps successfully?
Can your 9 year old baby sister or grandma follow the steps without asking for help?
- Can anyone raise a question in the github issues?
- Are you going to answer respectfully, short and concise and point the user to where to find the information, and take into consideration that perhaps something is not clear from your documentation and need more attention. (Perhaps the UI, a dependency or link has changed?)
- Or are you going to be a cocky ass and give degrading comments?
IF YOU WRITE IT GOOD, THEY WILL COME!
Now, available here:
- Don’t assume any previous knowledge!