Translate supported speech directly into English on device, keep a private transcript history, and move more cleanly between full transcription, live dictation, results, models, and subtitle workflows. This update also improves interface text scaling, light appearance, share/keyboard/widget integration, readiness feedback, and iPhone/iPad layout polish.

Import playable videos from Photos, Files, or the share sheet and transcribe their audio locally. Subtitle Studio can now render a new H.264 video with customizable captions, word-level karaoke timing, and distinct speaker colors or supplied speaker names. This update also adds low-latency live dictation and a custom keyboard, local noise reduction for the real-time path, longer sessions, clearer model readiness, and a full-quality large-v3-turbo lane for ordinary transcription.