Id like to have the best of both worlds, keeping things more disconnected when possible, but say “Hey Mycroft. Ask Google to <something Google assistant can do” when all else fails. I think it would ease adoption of the platform significantly since you could continue using your existing voice control as necessary. Does such a thing exist?
Not yet.
Two things, though: first you’re going to trigger google anyways sometimes with that command, and secondly, what are you trying to do on google that you can’t [yet] on mycroft?
I’ve thought a few times about turning off other assistants’ wake words, and trying to have Mycroft activate them and pass STT through. The problem is that, even if you can do it silently, Mycroft probably can’t pass the STT results through any faster than realtime, because it would just be translating it into some other Google/Siri/Alexa-audible frequency.
If you could dial into the other assistants directly, maybe…
The desire is more as a fallback. Transitioning to such things is usually not a one-to-one feature set, or it takes time to reconfigure. It’s part of why I haven’t done so yet. I suppose I could keep them both on, but it’s rather inelegantm
As for how, it would likely be a piece of code that would interact with the API directly that a number of devices support.
If you have the ability to install arbitrary Python on the device, and the name-brand assistant on the device has an accessible API, it should be possible in the long run to make a HiveMind or mesh node out of the thing - a nascent community project - which might do this. I doubt you’d be able to run Mycroft natively.
As for the how, a Mycroft skill that responds to “OK Google” would be easy. "OK Google " --> feed to the Google Assistant API. That part’s straightforward, once HiveMind gets the device-side part working well.
Can you point at the relevant APIs? I don’t have a Google Assistant device to mess with.