Let's clarify where these various nodes capture their audio.
node-red-contrib-mic can be used to access the microphone on the device running Node-RED - the audio capture happens in the runtime. That isn't going to work when running in the cloud.
node-red-contrib-browser-utils provides a node that lets you capture audio from the microphone on the device running the browser you have opened the node-red editor in.
Neither of these nodes can be used to capture audio from a node-red-dashboard ui. To do that, you'd need to use a ui_template node to insert some custom javascript that can be used to do the audio capture when triggered. I'm sure someone must have done that - but I can't find an example at the moment.
Capturing audio in the browser is a bit fiddly, but very doable. You could use the browser-utils node as a guide for what you'd need to do - https://github.com/ibm-early-programs/node-red-contrib-browser-utils/blob/master/microphone/microphone.html
