Changelog

macOS

v1.5.0May 12, 2026

New

  • Redesigned Voice Typing HUD — real-time audio waveform with visual feedback for speaking, low audio, and silence
  • Silence detection — warns you immediately if no audio is detected, so you never lose a recording to a muted mic

Improvements

  • Smarter AirPods handling — Bluetooth audio switching is now automatic and reliable across recording modes
  • Fn key dictation no longer pastes empty results into your text
  • Meeting transcript panel uses translucent background with quick-access stop button
v1.4.9May 8, 2026

New

  • Meeting recording — captures system audio and microphone simultaneously with live transcription every 30 seconds. Each meeting is saved as a self-contained folder for easy backup
  • Speaker-labeled transcripts — colored labels distinguish your voice from others. Tap any segment to jump to that moment in the audio
  • YouTube and video recording — automatically extracts the video title and produces a clean timestamped transcript with SRT export

Improvements

  • Better meeting detection — added support for Microsoft Teams, WebEx, Slack, Discord, and more
  • Meeting recording now only requires Screen Recording permission
  • Floating recorder no longer steals focus from your meeting app

Fixes

  • Fixed Google Meet falsely triggering meeting detection on non-meeting pages
v1.4.8May 6, 2026

New

  • SenseVoice speech model — 5× faster than previous CJK model for Chinese, Japanese, Korean with GPU acceleration. Download in Settings → Models (~830 MB)

Fixes

  • Fixed Fn quick dictation producing stray dots at the beginning of text
  • Fixed words occasionally getting cut off at the end of Fn recordings
  • Fixed long recordings sometimes appearing out of order
  • SenseVoice now correctly outputs punctuation
v1.4.7April 24, 2026

New

  • Qwen3-ASR speech model — optimized for Chinese, Japanese, Korean, and 27 other languages. Download in Settings → Models (~1 GB)

Improvements

  • Parakeet transcription is ~2.5× faster on long recordings
  • Qwen3-ASR supports streaming transcription — text appears chunk by chunk as you speak
v1.4.6April 23, 2026

New

  • Resizable AI Chat panel — drag the edge to adjust width

Improvements

  • AI polish is now on-demand — tap to clean up instead of auto-running
  • Default AI model upgraded from Gemma 2B to 4B — smarter responses overall
v1.4.5April 22, 2026

New

  • Local AI features powered by on-device Gemma 4 (Beta) — runs 100% on your Mac, your data never leaves the device
  • Auto-generated titles for long recordings
  • AI polish for Fn quick dictation — automatic punctuation, grammar, and casing
  • AI Assistant panel — ask questions about any transcript
  • Search now includes AI-generated titles

Other

  • AI features are in Beta and may change in future releases
v1.4.4April 21, 2026

New

  • Inline transcript editing — tap any paragraph in the detailed view to fix mistakes in place, no re-recording
  • “Clean up filler words” toggle — automatically removes um/uh/嗯/啊 during transcription

Improvements

  • Fn quick dictation starts noticeably faster, especially with AirPods and other Bluetooth audio
  • Smoother transition when transcription completes (no more bubble text-size jitter)
  • Better recovery from GPU hiccups during long meetings
  • Parakeet backend transcription quality improvements

Fixes

  • Minor polish and stability fixes across the app
v1.4.3April 15, 2026

Improvements

  • Noticeably faster Whisper transcription on long audio
  • Dramatically better recognition of proper nouns and specialized vocabulary
  • General pipeline stability improvements

Fixes

  • Major Whisper quality fix for long recordings — eliminated repetitive loops and dropped segments in meetings, interviews, and lectures
v1.4.2April 4, 2026

New

  • Playback speed control for recordings (0.5x–2x)
  • Recording auto-stop
  • Toast notification when your audio input device changes

Improvements

  • 5–10 seconds faster per transcription on average

Fixes

  • Fixed Traditional Chinese text rendering
v1.4.1March 21, 2026

Improvements

  • Redesigned Menu Bar — cleaner native menu replaces the old popover, with quick access to Settings and feedback
  • Improved shortcut hints — clearer guidance below shortcut buttons

Fixes

  • Settings window no longer flickers when reopened; it now centers over the main window automatically
  • Fn dictation and global shortcut can no longer interfere with each other — if one is recording, the other is safely ignored until the recording finishes
  • Fixed a launch issue where the app could start with a blank UI when set as a login item with the Dock icon hidden

Other

  • Localization improvements across 30+ languages
v1.4.0March 20, 2026

Fixes

  • Fn dictation no longer cuts off the last word — we now capture a short tail after you release the key, so final syllables come through intact
  • Fixed text insertion on non-QWERTY keyboard layouts (Dvorak, Colemak, AZERTY, etc.) — transcribed text now inserts correctly regardless of your layout
v1.3.9March 19, 2026

Fixes

  • Rapid Fn press/release no longer leaves the floating window stuck on “Transcribing…”
  • Fn+Backspace (forward delete), Fn+Arrow keys, and other Fn combos no longer trigger recording
  • Trial banner in sidebar now disappears immediately after license activation — no restart needed
  • Dock icon no longer briefly appears after reboot
  • Launch at Login no longer silently loses registration — the app re-registers automatically on startup
v1.3.8March 17, 2026

Improvements

  • Long Parakeet transcriptions now include automatic paragraph breaks at sentence boundaries
  • Restored completion sound effect when Fn key recording finishes

Other

  • Removed AppleScript paste option — keycode-based typing is now keyboard-layout-independent
v1.3.7March 14, 2026

New

  • New low-latency recording engine — rebuilt for faster start and more reliable audio capture
  • Seamless microphone switching — plug in or unplug USB/Bluetooth microphones while recording without interruption

Improvements

  • Automatically selects external microphone when connected
  • Fn key recording now briefly delays muting to avoid clipping the start of your speech, and no longer plays a sound when you stop

Fixes

  • Fixed text being inserted at the wrong cursor position when using global shortcut
  • Fixed recording getting stuck when unplugging USB microphone mid-session
v1.3.6March 13, 2026

New

  • Global shortcut now auto-inserts transcribed text at the cursor position in any app
  • Fn key recording automatically mutes system audio to avoid feedback

Improvements

  • Custom vocabulary UI improvements

Fixes

  • Fixed transcription stuck on “Transcribing…” when switching from Parakeet to Whisper
  • Fixed external keyboard users getting stuck at the Fn key test during onboarding
  • Fixed recording failure with some USB microphones — automatically falls back to compatible format
  • Fixed error message appearing when cancelling a model download
  • Fixed unnecessary microphone permission dialog appearing on some launches
v1.3.4March 7, 2026

New

  • Free trial with 5,000 words — try all features before purchasing
  • App Store users can migrate their purchase to the direct version

Fixes

  • Fixed keychain prompt appearing after upgrade
v1.3.3March 7, 2026

New

  • Automatic microphone switching — detects when you plug in or unplug a microphone

Fixes

  • Fixed sound effects being too quiet at low system volume
  • Fixed auto-update not working for some users
v1.3.2March 4, 2026

New

  • On-device translation — translate transcriptions directly on your Mac using Apple Translation framework, no internet required (macOS 15+)
  • Parakeet V3 speech engine — a new ASR engine is now built-in as the default, faster and more accurate for English

Improvements

  • Whisper Small model changed to downloadable on-demand, reducing app bundle size
  • Parakeet segmentation splits on sentence-ending punctuation for more natural paragraphs
  • Translation paragraph alignment is more robust, no more drifting between original and translated text

Fixes

  • Fixed transcription state cleanup to prevent rare “stuck” states after errors
  • Fixed translation map misalignment for older messages
  • Cached timestamps are now reused instead of re-transcribing audio
  • Model deletion now correctly resets the active model selection
v1.3.1March 4, 2026

New

  • Auto-update support — the app now checks for updates automatically via Sparkle
  • iCloud file import — audio and video files stored in iCloud Drive are auto-downloaded before import
  • Sidebar feedback button for quick access to support email

Fixes

  • Fixed app occasionally terminating unexpectedly due to watchdog firing during recording
  • Fixed iCloud import failure when file download status is unavailable
  • Fixed missing theme in sidebar Settings navigation
  • Increased model download timeout from 60s to 300s for slow connections

Other

  • Comprehensive localization audit and fixes across 32 languages
  • Removed PostHog analytics

iOS

For iOS release history, visit the App Store.