How to run other applications accessing sound devices?

[TheOldPresbyope]

How did you guess I was going to use “Computer” as my wake word?

https://youtu.be/hShY6xZWVGE



Actually, I don’t want any audio responses so a primitive moOde controller looks dead easy. If OTOH audio responses were required, the voice of the Star Trek computer (Madge Barrett-Roddenberry) was made available as a phoneme library several years ago.

[moode_user]

It would make sense to have this on the same Rasperry PIs as they are positioned well and would make a great
assistant.

I agree.

Snips.ai is a great voice assistant (with one of its advantages being that it can be run entirely offline) and since Moode clients are already connected to the audio system, it seems only natural to run Snips.ai and Moode on the same devices.

I currently run Snips.ai on a Raspberry 3B+ and Moode on separate Raspberry Pi Zero Ws in various rooms. 

In my opinion, a great setup would be Moode and a Snips.ai base on the Raspberry Pi 3B+ in one room and Raspberry Pi Zero Ws running Moode and acting as Snips.ai satellites in the other rooms.  An action would then be required to also be able to voice control Moode with Snips.ai.

Imagine this:

"Hey Moode, play Purple Rain by Prince". Moode has Spotify configured as its primary music source and plays Purple Rain from Spotify in the room where the voice command was received.

"Hey Moode, start with Let's Dance by David Bowie and play it everywhere". Moode starts the requested Spotify song radio and plays it on all Moode clients across the house.

Think about how awesome it would be for all parties involved if the Moode developer was able to team up with the people at Snips.ai and the librespot developers to make that happen.

[TheOldPresbyope]

@moode_user

From my exploration of snips.ai technology, I don't actually see much here that the moOde developer has to do. There are enough hooks already to get started and this would be just another item on his long ToDo list. 

This stuff is all open-source code. Instead of

Think about how awesome it would be for all parties involved if the Moode developer was able to team up with the people at Snips.ai and the librespot developers to make that happen.

think about how awesome it would be if someone like you did the grunt work to integrate a voice assistant with moOde, posted the results to github, and created a pull request for Tim (and the moOde test team) to consider. 

Just my 2-cents worth.

Regards,
Kent

[moode_user]

think about how awesome it would be if someone like you did the grunt work to integrate a voice assistant with moOde, posted the results to github, and created a pull request for Tim (and the moOde test team) to consider.

Ah, the good old friendly "it's open source, why don't you do it yourself instead of telling others what to do".

I am not a programmer. Basic things here and there I can code myself. I have done that in the past and shared the results with others. I certainly do not have the programming skills, though, to grab the moOde source code and write a suitable action in order to integrate Snips.ai.

Tim, the Snips.ai team and the librespot developers know their code inside out. If they like the "vision" that I shared, they are certainly the most qualified people to implement it with the least effort possible (in relative terms). Sharing the idea and bringing into discussion that they collaborate to realise it (provided they like it), is a perfectly reasonable thing to do and not an obscene act for which I need to be reprimanded, in my opinion.

If they do not like the idea or have other priorities or believe this would be as big of a task for them as it would be for a complete outsider who would need to familiarise himself with both code bases first - well, so be it. That is not my decision to make. I can offer to help with localising intentions, for example. Not enough? Ok. I am fine with that.

[Tim Curtis]

Hi,

Generally, major Open Source projects like Snips have huge teams, a lot of funding ($25M so far) but they are mainly focused on developing, supporting and promoting the Platform itself and it's API's. They don't do too much downstream integration unless its for a major client. They certainly wouldn't be interested in devoting a dev to moOde.
https://medium.com/snips-ai/inside-snips...f302364b44

For tiny project teams like moOde, we need to rely on the contributions of other developers to help integrate components like Snips.

I think it would be a great opportunity for someone and I know of at least one vendor that has a prototype HAT audio board with built-in mic's so I'd expect voice tech to start making its way into Pi audio moving forward.

Can anyone explain the licensing terms for Snips?

-Tim