The hardware is rather straightforward, something like a Raspberry Pi with off-the-shelf mems microphone module, powered Bluetooth receiver, battery + management (if you want it portable), and potentially some UART/SPI interface for more specific sensors/comms. The software I don’t have any direct experience with, but potentially something like Ada from HomeAssistant would be functional and easy to implement/integrate.
Is there a particular part of this that seems most challenging for you? Would you be interested in developing something to fill this potential gap? I’m toying with a lot of the same tech right now, I’d be glad to contribute.
This is the closest I have found, unfortunately, devices have been out of stock in many stores: https://github.com/espressif/esp-box
I think you can use a Bluetooth conference speaker for the audio part, they now have multi mic arrangements, echo cancellation, etc.
I created Willow, which uses the Espressif ESP BOX:
Boombeastic Mini’s Pi case design could be an option?
https://mycroft.ai/product/mark-ii/
Unfortunately it isn't going as well as they'd hoped: https://mycroft.ai/blog/update-from-the-ceo-part-1/
Or you could try reSpeaker (different options available): https://wiki.seeedstudio.com/ReSpeaker/
An older interesting comparison of mic arrays: https://medium.com/snips-ai/benchmarking-microphone-arrays-r...