Build an open future with us.

Invest in Mycroft and become a community partner.

Niceties and annoyances


#1

Would it be possible for Mycroft to automatically be in listening mode (as if the wake word had been spoken) whenever it’s speaking? That way, if Mycroft starts speaking - and you didn’t intend for that - you could simply say, “Cancel” rather than “Hey Mycroft - cancel”. It’s a little thing and might seem trivial, but to me, it’s way more usable that way. Seems like it should be the default behavior. Maybe make it configurable in settings. Maybe even set it on a skill-by-skill basis.

Along the same lines, could Mycroft automatically go into listen mode for a few seconds after each time it speaks? Again, it’s a simple/silly little thing but it would be nice if Mycroft said something to me and I replied, “Thank you” (as I tend to do with Alexa) and Mycroft responded with, “You’re welcome.”


#2

Hi there @DonnyBahama, thanks so much for your well considered feedback and suggestions.

Always listening when speaking

Unfortunately this is a limitation with some of our existing hardware, particularly the ARM-based Mark 1 and Picroft (which is based on RPi3 / RPi 3B+). This hardware doesn’t have any built-in noise cancellation, so if it were listening while speaking, then it would also “hear” what it was speaking, which in turns mean that the speech-to-text transcription would be likely not what the user intended.

We will have more options for this using the FPGA-based Xilinx processor in the Mark II. [See this link for all our Mark 2 updates]https://mycroft.ai/blog/tag/mark-2/().

Listening after speaking

This is a function that may be utilized in individual Skills. For example, if this is used in a Skill, then Mycroft will wait for a response and can further use the response in a Skill - such as providing a greeting.

See https://mycroft-core.readthedocs.io/en/stable/source/mycroft.html?highlight=response#mycroft.MycroftSkill.get_response