Meeting Mode now really identifies speakers. Record a conversation in Meeting mode and the transcript is split by who spoke, with each turn timestamped — then draft minutes from it Speaker turns are worked out on your device while you record. No recording is kept and no audio leaves the phone; only the labels are saved with the note Rename a speaker once and every one of their turns updates The first meeting you label downloads a one-time speech model (about 95 MB) over Wi-Fi Live Translate (Pro) — translation under the live transcript as you speak A new live bar — switch between Note, Meeting, and Translate in one place Control Center — start Transcription, Meeting Mode, or Live Translate in one tap Better East Asia support — Simplified Chinese, Traditional Chinese (Taiwan), Japanese, Korean Optional cloud AI for China and nearby markets — configure Qwen or DeepSeek when Apple Intelligence isn’t available Transcription accuracy is still never held back for Pro

Apple Speech language packs finish setup reliably — download no longer reports success while the language stays “not ready” Clear Activate step when a pack is downloaded but not yet active for this app Spotlight & Shortcuts: open and search your notes; start recording from Siri / Shortcuts Open a specific note from Spotlight, Shortcuts, or Siri Delayed Camera Actions: “in X minutes / hours” and clock times (not seconds-only) Includes 2.3.0: Whisper 1.9.4, Silero VAD, Long Meeting Mode (Pro) Transcription accuracy is still never held back for Pro

More reliable Whisper transcription (updated speech engine) Better speech / silence detection for live Auto Chapters Long Meeting Mode (Pro) — adaptive capture for longer Whisper live sessions Auto Chapters (Pro) still included — timed, jumpable sections after live capture Transcription accuracy is still never held back for Pro

Auto Chapters (Pro) — after a live recording, get timed sections you can jump to Chapter titles written with on-device AI (you can cancel while they generate) Edit the text inside each section More reliable on longer recordings Meeting Mode, Improve, and Format refresh chapters instead of leaving stale ones behind Stop → Skip / Save feels snappier (heavy finalize work deferred until you need it) Transcription accuracy is still never held back for Pro

More stable live transcription — fixes threading around Apple SpeechAnalyzer audio capture and bounds memory use on long sessions Apple Speech models are retained only while in use, then released after stop/cancel Home Screen widget to start recording with one tap Freer translation — pick from many languages (including Chinese, Korean, Japanese); live From locks to speech language; Notes can swap From/To Editable Actions — tweak Reminder / Calendar / Alarm title and time before adding Reliability and guideline polish (including speech-recognition usage description on the app target)

Unlimited private library with Pro — free keeps 20 notes on the phone Ask your notes with on-device Apple Intelligence Study dictionary that lights up words in your transcripts One-tap translate — free, always there Actions from a note → Reminders, Calendar, alarms, and Camera whisper.cpp 1.9.2 Pro is still a one-time purchase. Restore anytime. No subscription.

1. Apple Speech language packs now finish setting up. Downloading a language could previously get stuck reporting “not ready” forever — most visibly for Cantonese and Hong Kong Chinese, and for any language you added after the first 2. The app no longer quietly transcribes in a different language than the one you selected 3. You can now see how many Apple Speech languages are active and remove one to make room for another 4. When a language can’t be set up, the app explains why instead of showing the same generic warning

More reliable transcription — the speech engine now recovers on its own if it is interrupted while the app is in the background, instead of quietly returning empty transcripts until you restart Faster, more accurate summaries for Chinese, Japanese and Korean recordings Very short recordings are now handled gracefully

NEW in 1.2.0 — Meeting Mode! Turn a whole meeting or lecture into a clean summary with key points, decisions, and action items — entirely on-device, even for long recordings. Plus Voice Commands, categories, and favourites. Still 100% private.

- Voice Commands (reminders, URLs, clipboard) with one-tap execution - AI-powered command detection with offline regex fallback - Notes categories with custom icons and colours - Favourites for quick access - Clipboard commands - Simplified Chinese localisation — full app UI now available in 简体中文

A few improvements in this update: • Added a Copy and Share button in the AI tab so you can quickly save the AI's response • Fixed a bug with transcribing audio from imported audio files

Version 1.0.1 adds Apple Speech live transcription support on compatible devices, with bundled Whisper Large available as the on-device fallback engine. This update also improves AI polish features with Clean Up, Smart Notes, and Summary & Actions. When supported, AI polish can run with Apple Intelligence on-device; otherwise it can use your configured cloud AI connection.