Cliptate v1.2 · More transcription models, smoother dictation and editing

Features and improvements

• Transcription models: AI Cut and Voice To Text gain multilingual Whisper and Nemotron ASR, with shared model downloads and optional precise word alignment.
• CLI and history: Generate speech from text, query transcript history, and inspect saved AI corrections to improve vocabulary. History preserves correction evidence and the Restore Original baseline.
• Live dictation: Previews revise growing speech windows, streaming audio is more reliable, and new microphone indicator styles are available.
• Editing workspace: Unified sidebar and page layouts improve video import, playback, and navigation. Free builds remove the Free Access navigation entry.
• Updates and AI: The installer prompt shows verified update-signature status, and OpenAI requests omit unsupported generation parameters. Cliptate Dev separates its identity, settings, history, and CLI transport from production.

Upgrade notes

• Requires Apple Silicon and macOS 15+. Update in the app or close Cliptate and replace it using the DMG. Restart the app before using the upgraded CLI.
• Settings, history, compatible caches, and model selections are preserved. Download newly selected models first; old history gains no fabricated corrections. GitHub builds remain free and App Store rules are unchanged.

Known limitations

CLI transcription and speech generation still create no history entries. Nemotron live dictation uses the shared VAD pipeline without a native streaming encoder cache. Real-model inference, microphone input, and update installation still need manual validation.