Since I didn't manage to get the audio stream from the microphone, I made a local addon that runs this command:

ffmpeg -f alsa -ac 1 -i default -c:a libopus -ar:a 16000 -ac:a 1 -af adelay=0ms -rtsp_transport tcp -f rtsp rtsp://localhost:8554/webcam

updated: had pasted the wrong command

And I have a stream called webcam in the go2rtc config with only video. When I open it i the webrtcard, it magically picks up both video and audio. Is this the right way of doing it?
I also tried creating the final stream with both video and audio from my addon, but then I don't have where to send it to (since no stream is defined in go2rtc)

What would be the right way of doing this?

Btw, in my local addon I have these packages:

RUN  apk add --no-cache \
        alsa-plugins-pulse=1.2.7.1-r0 \
        alsa-utils=1.2.8-r0 \
        pulseaudio-utils=16.1-r6 \
        ffmpeg 

and in the config.yaml, I added:

audio: true
video: true

Maybe adding alsa and pulseaudio to the go2rtc docker container would allow for using the mic?