ws-audio-api | WebSocket Audio API : library to broadcast the sound | Audio Utils library

I am trying to make an as-simple-as-possible Javascript frontend that will allow me to receive audio from a user's mic on a mouse click within a web browser using getUserMedia, modify it to a custom sample rate and monochannel, and stream it over a websocket to my server where it will be relayed to Watson Speech API.

I have already built the websocket server using autobahn. I have been trying to make an updated client library drawing on whisper and ws-audio-api but both libraries seem outdated and include much functionality I don't need which I am trying to filter out. I am using XAudioJS to resample the audio.

My current progress is in this Codepen. I am stuck and having trouble finding more clear examples.

1. Both whisper and ws-audio-api initialize the AudioContext on page load, resulting in an error in at least Chrome and iOS as audio context must now be initialized as a response to user interaction. I have tried to move the AudioContext into the onClick event but this results in my having to click twice to begin streaming. I am currently using audio_context.resume() within the onClick event but this seems like a roundabout solution and results in the page showing it is always recording, even when it's not, which may make my users uneasy. How can I properly initiate the recording on click and terminate it on click?
2. I have updated from the deprecated Navigator.getUserMedia() to MediaDevices.getUserMedia() but not sure if I need to alter the legacy vendor prefixes on lines 83-86 to match the new function?
3. Most importantly, once I get a stream from getUserMedia, how can I properly resample it and forward it to the open websocket? I am a bit confused by the structure of bouncing the audio from node to node and I need help with lines 93-108.