Version 1.6.5
April 2, 2026
- Professional audio engine — Completely rewritten audio recording system for bulletproof microphone handling. AirPods, Bluetooth headsets, and USB microphones now switch seamlessly without freezing.
- Visual mode indicator — White pill when using Local Models, blue pill for Cloud AI. Always know which mode you're in at a glance.
- Automatic retry — If a server error occurs, Sosin retries automatically before showing an error.
- Stability improvements — Fixed menubar icon getting stuck, improved model warm-up messages, smoother cancel animation.
Version 1.6.4
April 2, 2026
- Media pause works with AirPlay, HomePod, and Bluetooth — Sosin now pauses Apple Music and Spotify during dictation — including when playing through external speakers. Music resumes automatically when dictation ends.
- Seamless audio output switching — Freely switch between built-in speakers, AirPlay, HomePod, and Bluetooth in Control Center — Sosin handles it seamlessly.
- Improved non-English language detection — More accurate language recognition for non-English dictation, especially for shorter recordings.
Version 1.6.3
April 1, 2026
- Bug fixes — Minor stability and compatibility improvements.
Version 1.6.2
April 1, 2026
- Improved AI transcription reliability — Longer dictations are now processed more reliably. Reduced phantom words at the end of transcriptions.
- AirPlay & HomePod media pause — Media pause now works with AirPlay speakers and HomePod. Grant Sosin access in System Settings → Privacy & Security → Automation.
- Stability improvements — Various performance and reliability improvements under the hood.
Version 1.6.1
March 31, 2026
- Bug fixes — Improved handling of mixed-language dictation and minor stability improvements.
Version 1.6.0
March 31, 2026
- Faster AI processing — Dictation results now appear almost instantly — significantly faster than before.
- Bug fixes — Parallel dictation now pastes results in recording order. Fixed cost estimate for BYO Key users. Improved model warm-up messages. Windows now open in correct order.
Version 1.5.4
March 29, 2026
- Parallel dictation order fix — When dictating multiple messages in a row, results now always paste in the order you recorded them — even if a shorter dictation finishes first.
- Improved model warm-up — Switching back to Local Models now shows a clearer warm-up message instead of the first-time activation text.
Version 1.5.3
March 28, 2026
- Bug fixes — Various stability and animation improvements.
Version 1.5.2
March 27, 2026
- New pill animation — The completion animation now uses the same visual language as recording — bars flatten to a line, then collapse to a dot. Inspired by classic CRT TV shutdown.
- Bug fixes and improvements — Various stability and UI improvements.
Version 1.5.1
March 27, 2026
Anonymous error reporting — Help us improve Sosin by opting in to anonymous error reports. No personal data is collected. Enable in Settings or during setup. Thank you to everyone who turns this on — it makes a real difference.
Bug fixes and improvements.
Version 1.5.0
March 27, 2026
- Parallel dictation — Start a new recording while the previous one is still processing. Both texts appear in order.
- Apple AI model — On-device transcription powered by Apple Intelligence. English only. No download needed.
Version 1.4.1
March 26, 2026
Mute audio fix: no longer starts paused media. Short recordings silently discarded instead of showing error. Faster repeated dictation.
Version 1.4.0
March 26, 2026
- Paste Last Dictation — Double-tap Option to re-insert your last dictation when the wrong window was focused. The key auto-pairs with your dictation hotkey — always on the same side of the keyboard.
- Pause media during dictation — Music and video automatically pause when you start dictating and resume when you stop. Works with Spotify, Apple Music, YouTube, and more. Toggle in Settings.
Version 1.3.0
March 22, 2026
- Local Models — Offline dictation with on-device AI — no internet needed. Download a model once, dictate anywhere.
Version 1.0.0
March 17, 2026
Initial release.