Select devices
- Open Settings → Devices.
- Choose Microphone for Voice mode and the composer microphone.
- Choose Speaker for spoken replies, or System default to follow the OS selection.
- Save settings and test the intended feature with a short input.
If a saved choice appears as Saved device (disconnected), reconnect it or select an available device. Device labels can depend on microphone permission. Speaker selection support can differ by OS/runtime, so confirm the actual output rather than relying only on the dropdown label.
Configure sounds and notifications
Open Settings → Sounds. Use Interface sounds and Interface sound volume for short task, approval, message, and voice cues. Muting these cues does not necessarily mute a voice-model response or change the OS output volume.
Under Event mapping, choose the sound assigned to each event. Moving through a sound picker previews choices; the labeled event button plays the event cue. Desktop notification controls separately cover task completion, failure, and approval-needed events.
Distinguish hardware from providers
A working microphone does not prove the transcription or live voice model is configured. If recording works but transcription fails, inspect Preferences → Dictation and its provider connection. If voice text appears but nothing plays, inspect Speaker, OS volume, and the speech route.
Use voice/audio recovery, Voice, and Dictation and read aloud. Native device selection and permission prompts were not exercised for this draft.