This archive summarizes stable releases using the dates in their release documents. Current source supports free direct downloads only; historical App Store and subscription references describe those releases. Nightly previews are separate. See the English source for commits and validation.
Features and improvements
- Native CLI: Install from Settings → Command Line, batch-transcribe audio/video to text, JSON, Markdown, SRT, or VTT, and manage custom vocabulary.
- Dictation reliability: Clearer microphone diagnostics, better handling of trailing speech, and saved live transcripts when final recognition fails.
- App startup: Added launch-at-login settings and moved system queries off the UI thread. CLI launches run in the background without opening the main window.
- Model compatibility: Updated MLX inference dependencies improve transcription, diarization, and speech synthesis compatibility.
Upgrade notes
- Requires Apple Silicon and macOS 15+. Update in the app or close it and replace via DMG; verify its SHA-256 file. Move the app into Applications before installing the CLI, add ~/.local/bin to PATH, and restart after upgrading.
- Direct GitHub builds now run without App Sandbox. Missing settings, presets, and history are copied from the old sandbox, preserving originals and existing direct-build data. Compatible Hugging Face caches are reused.
Known limitations
The CLI accepts local files, processes one transcription at a time, and creates no projects or history entries. App Store builds retain sandbox, subscription, and Accessibility restrictions and do not include the CLI.
Original release on GitHub
This archive summarizes stable releases using the dates in their release documents. Current source supports free direct downloads only; historical App Store and subscription references describe those releases. Nightly previews are separate. See the English source for commits and validation.