v1.5.1

Release v1.5.1

v1.5.0

What's Changed

• Added Windows CUDA 13 builds to support GPU acceleration on newer graphics cards

v1.4.0

New Features

• Voice Activity Detection (VAD)

v1.3.0

What's Changed

• Quick search: Search transcript text as you type
• Clear transcript: Click the 'x' next to the Transcript heading to clear the transcript and start over
• ARA persistence: Faster, event-driven transcript storage using ARA (see note)
• Audio source selection: Better visibility of audio sources to be processed, with ability to select and deselect

Note on ARA persistence

This release will migrate transcripts saved by older versions. The ARA persistence change is backward-compatible (newer versions of the plugin can read older versions' data), but once saved, older versions of the plugin will not be able to read the transcript.

Additional notes

• Transcripts in multiple instances of the plugin will now stay in sync with each other
• Linux CUDA builds are temporarily disabled due to build worker disk space issues

v1.2.2

What's Changed

• Remove CUBLAS DLL from Windows builds

v1.2.1

What's Changed

• Update JUCE to version 8.0.8
• Fix bug with stuck progress bar

v1.2.0

What's Changed

• CSV and SRT export
• Better pause handling

v1.1.2

What's Changed

• Crash fix: Fixes whisper.cpp initialization bug

v1.1.1

What's Changed

• Build for all CUDA architectures: Support all available GPUs when using CUDA

v1.1.0

What's Changed

• Follow playback: The transcript will now highlight and scroll to the speech that is currently playing
• Download progress: The progress bar will show the progress of speech recognition model downloads
• Cancellation: While processing, the "Process" button will turn into a "Cancel" button that can be used to cancel transcription