Speech Settings
Tap the More Menu in the upper-right corner and choose Settings, then select Speech.
Speech Engine
Choose the speech-to-text engine Action Phrase uses while listening.
- Apple Speech
- FluidAudio
Switching engines changes which Speech options are shown below.
FluidAudio
When Speech Engine is set to FluidAudio, these settings are available.
Diarization
Choose how speaker handling works:
- SpeakerManager
- Sortformer
- LS-EEND
Model Version
Choose the FluidAudio ASR model version:
- V2
- V3
Changing model version requires model files to be downloaded again.
Voice Enrollment
Opens the Voice Enrollment screen.\ This row appears for supported diarization modes and shows current status (for example None, Auto, or a selected profile).
Silence Detection
Controls how quickly speech is considered complete after silence.
Options:
- 0.4s
- 0.6s
- 0.8s
- 1.0s
- 1.5s
- 2.0s
- Disabled
Advanced FluidAudio
Opens a dedicated Advanced FluidAudio settings page.
Apple Speech
When Speech Engine is set to Apple Speech, these settings are available.
Language
Choose System Language or select a specific supported language/locale.
Debounce Partial Transcript
Controls how long Action Phrase waits for the transcript to stop changing before evaluating phrases.
Options:
- Disabled
- 1 second
- 1.5 seconds
- 2 seconds
- 5 seconds