Read-aloud translation Plug in headphones and Sonasi reads the translation aloud while you record, and stops when you unplug. New on-device voice, Sonasi TTS (Vietnamese): works offline, 25 voices. Choose your headphones, volume and left or right ear. If a cloud voice fails, it switches to an on-device voice.

• When an older recording can't be decrypted, the app now tells you why instead of showing a dead player, and also tries the keys used by earlier versions. • Fixed how existing users are recognized after reinstalling the app or moving to a new Mac, so lifetime Pro is kept correctly. • Bug fixes and stability improvements.

Fixed several UI issues.

What’s New More languages, on device Live transcription now runs on device for many more languages, including Vietnamese, Thai, Hindi, Russian, and Indonesian — no cloud needed. When your language is high quality on Apple’s on-device engine, cloud transcription automatically switches to on-device to cut cost and latency, with no loss in quality. Better live accuracy Fixed dropped and out-of-order words during live transcription. Rebuilt real-time cloud transcription (ElevenLabs and OpenAI): no more cut-off, duplicated, or vanishing lines. Auto-detect source language now translates correctly. Redesigned floating captions (Picture-in-Picture) Text now streams in live, word by word, instead of jumping a whole sentence at a time. Easier to read: the translation stands out, the original sits beneath, every line is timestamped. A fixed, resizable window — captions scroll inside instead of growing off-screen — and it stays in sync with the main window. A window that fits your screen The main window now resizes much smaller, down to a compact size that’s handy during meetings. New compact button (next to Picture-in-Picture) instantly shrinks the window and brings it back to full size. The transcript is now one clean, easy-to-read stream at every size. Your choices, remembered The app remembers your Live display mode (Transcript only, Bilingual, or Translation only) and your language selection. Clearer guidance Friendly alerts when a language may be low quality on the default engine, or when a language pair isn’t supported — with one tap to switch to Cloud or Local, or download the language pack in app. Privacy and sharing Encryption now protects your audio file only; transcripts are always searchable in Sonasi Notebook. New installs start with Sonasi Notebook sharing on and audio encryption off (both changeable in Settings). Optional anonymous usage statistics (opt-in during setup) help us improve the app — never your audio, transcripts, or any personal data. Also: quick links to the other Sonasi apps and a “Report a problem” page, plus a range of stability and performance improvements.

Version 1.1 — much better recognition & translation • On-device (Default) speech recognition now supports VIETNAMESE and 20+ more languages — runs locally, no network needed. • Live Cloud captions are far smoother and more accurate: no more mid-sentence cuts, dropped words, or duplicated/vanishing lines. • Fixed translation not working when the source was set to Auto-detect. • Remembers your choices: source/target language, Interpreter, and view mode (Original / Bilingual / Translation only). • Clear warnings when a language or translation pair doesn't suit the chosen engine, with one-tap switch to Cloud/Local and in-app language-pack download. • Save recordings to Sonasi Notebook to search transcripts (on-device; audio never leaves Meetings). • Various recording and recognition stability fixes.

Allow Sonasi Meetings to sync data directly with Sonasi Notebook, another application within Sonasi AI Studio.