Mark 2 Ideation


We are curious about your thought, ideas and opinion about Mark 2.
What should Mark 1 become? What do you like about it? What would you change? What can we do to serve you better?
Please share with us your thoughts about the future of voice assistants and AI in general!

Aura from the Mycroft AI Design team.


Barge in ability with hardware. Meaning, when something is playing I can interrupt it and ask for a new skill.

Ability to hear about cool new skills by activating some sort of skill like “trending skills”

Teaching Mycroft to be smarter skill. Basically, let me train the AI from the unit.

Longer reach of mic. I want to be able to sit in a room that is large and speak freely to Mycroft.


Improve microphone quality/range.

The Mark 1 rubber feet are a bit loose.

The rotating button on the Mark 1 is a nice, simple interface, but some of them haven’t worked quite right. Improve reliability.

Perhaps a way for the face animations and information that is currently displayed on the faceplate to be shown on an external screen. This would make it easier to display more information and/or pictures/video directly without needing a built-in screen on the unit (which would increase cost).


Perhaps two types of units? A smaller unit that’s only got indicator lights, mic array, wifi (call it a sherlock, maybe). The Mycroft 2 should have a camera, and an (5-10"?), and more or better mics. It would be amazing to see an option for local voice processing, or an option to add it in somehow. Unlikely at the current moment I know.

  • Improve the mic. The current mic generates some white noise and you have to get quite closer to mark1 to be able to understand what are you saying.

  • Include an external camera. Maybe rasberry pi camera?, it should be easy to implement (I’m trying to do the same with my current mark1 prototype). Then it would be possible to include some extra skills like “take a picture”, “detect movements” or a facial recognition system with opencv.


Love the idea of Mycroft and Sherlock.


Accessory market is interesting. I also would like to upvote that option for the camera. You could really have fun with your pets with a camera and voice during the day. :slight_smile:


Yes!! more ways to custom your capabilities can just increase the fun! “capturing moments in your home” type of scenarios for pets maybe “the patio light turns on when my god passes throughout at night”


Some thoughts off the top of my head:

  • No more wall warts! Put the brick in the middle of the cord if you must, but electrical outlets are prime real estate. Don’t be greedy.
  • an ON button (or in-line switch on the cord)
  • Bluetooth headset or other way of providing voice commands at a distance
  • respond to and/or generate DTMF or other tone sequences:
  • “pass code” unlocking
  • better interpretation in a noisy environment with multiple conversations
  • custom alarms / alerts
  • wall mount (with a different orientation for the “face plate” / visual status indicators)
  • choices for case color
  • Wacky thought: A Rust-based OS when someone gets one written. :wink:


Mycroft must have a British accent obviously!!


Of course, good news for you! We are currently working on our American Female, American Male and… British Female voice! Eventually we will also have a British Male voice.

Our American Female voice is going to be released first and her name is Trinity. :raised_hands: :slight_smile:


Question of the day!
What are the TOP 3 things your Open Source Assistant needs to do very well?


Hey all, I can tell you we are definitely working on improving the following:

Microphone quality
Sound quality
Barge-in capability

Thanks for the suggestions. Accessories, wall mounts, color options, a smaller unit are all great ideas!


Honestly, I really like the Mark 1 and all I really want is some refinement. Top wish being a better and longer range microphone. Better audio quality/volume. That’s about it for he hardware.

To be extra honest. I don’t really care if you do anything different other than a better mic. What would really be the best is throwing everything you can at improving the software and increasing the number of skills.


I should be able to change the name easily to whatever I want to call it.

If you can integrate SAPI voices, I already own some.

Offline voice recognition is imperative. Google made a proof of concept that’s only 26MB… Ask me for the link to the article…

Here is your edge: IR and RF receivers and transmitters - so Mycroft can take command of my existing dumb appliances.

Get Bluetooth going… Multipoint Bluetooth. Multi, multi multi point… Bluetooth speakers, Bluetooth mics… many devices simultaneously… Make it a Bluetooth hub… All my home automation tinkering so far has been with Bluetooth not IOT and I’ve done that because IOT can be hacked from the other side of the world, Bluetooth can only be hacked within Bluetooth range…


I like the idea of being able to tell mycroft to turn my tv on and change the input.


Controlling stuff via Bluetooth and IR are great ideas!


@steve.penrod @forslund

Hope yo see this. Not sure exactly what the SAPI voices are, but interesting as you guys are working on voices.

I also agree with the BT hub.


Can I get said link to that article?


Thanks for the input everyone. From my perspective:

  1. Array microphone with beam forming and barge in capability.
  2. Better speaker quality ( proper volume for speaker selection ).
  3. Screen.

I’ve given these requests to the design team ( Derick, David and Aura ) and am really excited to see what they come up with.