What’s New
-
On-device transcription with FluidAudio
FluidAudio is now the default speech-to-text option. Models download on first use and run locally on your Mac. -
Speaker identification
VibeScribe can now detect different speakers, add speaker labels to the transcript, show a speaker timeline, and let you rename or merge speakers. -
App language setting
The interface is now available in English, German, Russian, and Simplified Chinese. -
Audio export
You can now export the original audio file from the record actions menu.
Improvements
-
Better system audio recording
System audio capture and permission handling are more reliable. VibeScribe now asks for system-audio access only when needed. -
Cleaner transcript formatting
Timestamped transcripts are easier to read, with better handling of split segments and missing tail text. -
Better Whisper-compatible API support
Improved compatibility and fallback behavior for external Whisper-compatible transcription servers. -
Polished record details
Copying transcripts and summaries is clearer, actions are more consistent, and tag handling is smoother. -
Better recording overlay and settings
The floating recording overlay and the Settings screen now feel more consistent and easier to use.