Mute system audio output during dictation (for live calls like Discord/Zoom), not just MediaRemote pause

[kskaletski]

Problem

The existing "Pause media while dictating" option only pauses apps that expose a Now Playing / MediaRemote session, such as Apple Music, Spotify, and YouTube. It does not silence live communication apps like Discord, Zoom, Google Meet, or FaceTime, because they do not publish MediaRemote state and do not respond to a pause command.

When I dictate during a Discord call, the other participants' voices keep playing through my speakers, and the microphone picks them back up, so their speech gets transcribed into my text. In practice, FluidVoice dictation is only reliable when I am alone in a quiet room.

Proposed behavior

Add an option that makes system audio output inaudible for the duration of a recording, and restores the previous state when the recording stops. The key requirement is the observable behavior: while I am dictating I should hear nothing from other apps (including live calls), so nothing leaks back into the mic, and normal sound returns the moment dictation ends.

This is the behavior I observed with Wispr Flow: pressing the dictation shortcut immediately silenced everything coming out of the Mac, I dictated in silence, and sound came back afterwards. I did not need to hear the call while dictating, which is exactly what I wanted.

I am describing the behavior, not prescribing an implementation. One possible approach could be muting or ducking the default output device via CoreAudio (for example kAudioDevicePropertyMute, or a save-and-restore of output volume), but the maintainer is best placed to pick the right mechanism.

Ideally this is a toggle separate from "Pause media while dictating", for example "Mute system audio while recording", since the two solve different cases and some users may want both.

Alternatives considered

• "Pause media while dictating": does not cover live communication apps, which is the core problem here.
• Manually muting the Mac before each dictation: works, but defeats the point of a fast push to talk flow.
• Using headphones: avoids re-capture, but I want to keep using speakers and simply not hear the call while dictating.

Environment

• FluidVoice on macOS, Apple Silicon
• Compared against Wispr Flow, which silenced system output during dictation