SuperWhisper News
Faster & improved language detection
October 13, 2026
Faster & improved language detection. Smarter timeout handling for S1. "Always Close" is now enabled by default. Fixes The agent chat no longer pops up over an agent session you're already in. Improved paste reliability in Chromium browsers by retrying when a paste is dropped. Fixed paste formatting in Signal Desktop and other Electron apps. 2.19.2
Fixed an issue where app updates could get stuck during installation
October 5, 2026
Fixed an issue where app updates could get stuck during installation. 2.19.1
Fixed a crash that could occur on a fresh install
October 5, 2026
Fixed a crash that could occur on a fresh install. 2.19.0
Faster microphone start when beginning recording across all input devices
October 2, 2026
Faster microphone start when beginning recording across all input devices. New Muffle sound option applies a low-pass filter to other apps' audio while you record. Lower playback now drops audio to 30% of your current volume instead of a fixed level, so it scales with how loud you're listening. Fixes Fixed issue with "Always Close" option when paste detection fails. Fixed license details not updating in the UI after unlinking a personal license. Fixed trial progress bars and trial cards not displaying correctly in the sidebar and License view. Fixed text flicker while typing in the agent composer and when sending messages in agent chat. Fixed clipboard restore issues when using "Keep what I have copied." Fixed the reprocess button unintentionally selecting a mode on a plain click. 2.18.5
Better filler word removal on Scribe realtime for cleaner transcriptions
September 28, 2026
Better filler word removal on Scribe realtime for cleaner transcriptions. Fixes Fixed real-time transcription dropping audio. Fix high CPU usage after playing or pausing a recording in History. Fixed CSV vocabulary import for files with Windows line endings. Fix to pin local models on Pro mode upgrade. 2.18.4
Bring-your-own-key (BYOK) setups now respect your organization's cloud model settings
September 23, 2026
Bring-your-own-key (BYOK) setups now respect your organization's cloud model settings. Improved handling of longer recordings with the S1-mini model for more reliable transcription. Fixes Pro-only features now initialize correctly only after Pro access is confirmed. 2.18.3
Silence removal is now enabled by default for cleaner transcriptions
September 3, 2026
Silence removal is now enabled by default for cleaner transcriptions. You can turn it off in Sound settings. Your clipboard contents are now preserved after dictation by default. Fixes Fixed an issue where redirected pastes could remain open when confirmation wasn't available. Fixed a crash in the agent editor on macOS 27. Fixed recorder window dragging on macOS 27. Many other macOS 27 fixes. Fixed hover interaction on the Mini Window when dismissing a recording. 2.18.2
All local Whisper models are now free on macOS, giving you high-quality voice to text at no cost
August 26, 2026
All local Whisper models are now free on macOS, giving you high-quality voice to text at no cost. Agent plugin support for DeepSeek Harness. Improvements to S1-mini consistency. Free Pro trial changed from 15 minutes to 3,000 words, and reset for existing users. 2.18.1
Updated default modes and models for a smoother onboarding experience
August 20, 2026
Updated default modes and models for a smoother onboarding experience. Improved S1-mini stability on Intel-based Macs. 2.18.0
Added S1-mini, an on-device language model with tone control
August 19, 2026
Added S1-mini, an on-device language model with tone control, that cleans up and formats your dictated text without a network request. Local Mode is now suggested automatically when your keyboard is set to English, making it easier to start with private, offline transcription. 2.17.3
Added tone control when using Message, Mail, and Notes, giving you more say over how your text sounds
August 17, 2026
Added tone control when using Message, Mail, and Notes, giving you more say over how your text sounds. Redesigned the model details sheet with a cleaner layout, download size info, and clearer confirmations for downloading and deleting models. Refined tooltips throughout the app for a smoother, more consistent experience. Improved language detection accuracy for more reliable transcriptions. Fixes Fixed a stuck "Discard recording?" prompt in mini mode. Resolved compatibility issues on the latest version of macOS. Stopped previously copied content from leaking into super mode results. Fixed an issue where transcription could be blocked, ensuring faster and more reliable results. Made shortcut hover interactions feel snappier and more responsive. 2.17.2
Introducing the S1 family of models
August 6, 2026
Introducing the S1 family of models. Improvements Faster, smoother settings window with refined picker, disclosure, and Appearance card animations. Short silent clips are now skipped before transcription for improved reliability. Fixes Cohere Transcribe now preloads from recording start so short transcription are fast from cold start. Fixed copy, undo, and clipped text issues in the agent prompt box. Corrected settings window placement. Fixed an issue where the wrong message could appear after advancing the agent message queue. Improved autocapitalization when inserting text after a newline. Fixed the X link in the license view. Resolved a model download reliability issue. 2.17.1
Automatic language detection for Cohere Transcribe Refreshed the app icon The app name is now capitalized to…
August 4, 2026
Automatic language detection for Cohere Transcribe Refreshed the app icon The app name is now capitalized to match updated branding. 2.17.0
Scribe v2 Realtime in superwhisper
July 6, 2026
superwhisper added ElevenLabs' Scribe v2 Realtime, a voice-to-text model that writes as fast as you talk with the accuracy to keep up. It supports 90+ languages with automatic detection, and you can mix languages within the same dictation. Available across Mac, Windows, and iOS.