@secemp9@BillyM2k You don't need to send audio anywhere, a local voice recognition engine could just spit out keywords in text format to match against available ads.
@skymen75@TylerGlaiel If you fork and diverge you will miss out on upstream changes and it might be hard keeping up with new features and improvements contributed by other devs.
I would make modules and tools on top of the engine in a separate repository, so you get the full benefit of foss.
@jaleelshaw I still buy records and merchandise of music I truly love. Even though I have a subscription. Would not buy a record for the 3 times I stream most songs. I use the service mainly for discovery.