Hey mycroft we have to talk: Talks without hey mycroft

People can talk without key words. I thought that would be a great selling point for mycroft. Here are my thoughts:

  • a request could start with “hey mycrft we have to start talking”.
  • “if you think” mycroft begins the audio recognition based on the level.
  • as long as volume is detected, and the server request starts.
  • during the playback of the answer, no detection is performed.
  • Now mycroft is waiting for audio again. if you’re say “great talking” the end of the recognition.

Much time could pass between inquiries. This is the mode when you alone in the room a book reads or you want to call several things in a row but takes time to think about the question

What do you think. would that be possible.

There’s the concept of a transaction within skills, ie, “hey mycroft turn on some music.”
and mycroft would ask what you want to hear/feel like listening to, and you respond with “my favorites playlist”.

There’d have to be a ridiculous decision tree attached to what you’re discussing to account for relevance, but not implausible.

I know the talk tree. However, I am always forced to an answer and I can not change the time for the answer.

Of course, listening permanently is not so great in privacy, but with homeserver no problem

hey lately i got deepspech to work on my server in german. then i started a similar dictation skill https://github.com/gras64/deepspeach-dictation-skill . I don’t have the ability to set up mycroft for permanent listening for a certain period of time. deepspech can currently run from the resources on a “normal pc”. does my skill have to connect the server directly or can i do it in the future via a mycroft function.