This update focuses mainly on improvements around responsiveness in STT→TTS mode.

Main Changes

・Improved response speed in STT→TTS mode ・Adjusted detection behavior to reduce false responses when no speech is present ・Reviewed handling of short utterances and pauses between speech ・Adjusted internal processing to make conversations feel smoother and more natural

Recommended Setting

After updating, we recommend resetting your speaker profile.

If you continue using a speaker profile created in an earlier version, some of the latest adjustments may not be fully reflected. If you use STT→TTS mode, please reset your speaker profile once and set it up again.

Updated the real-time transcription engine for improved stability and compatibility.
Added automatic recovery to a safer transcription engine when startup issues are detected.
Improved microphone selection and testing during first-time setup.
Improved fallback handling for systems where GPU/Vulkan acceleration is unavailable.
General crash-resilience improvements.