Since beta
Everything we've shipped from v0.6.3 until now.
Headliners
Windows (0.8.4)
Glimpse on Windows 10+.
Glimpse Personal (0.9.0)
Glimpse now has a paid tier for the advanced features. Dictation stays free, with a two-week trial on the rest.
Apple Neural Engine (0.9.6)
Whisper now runs on the Neural Engine on Apple Silicon, making transcription faster and more power-efficient.
Import from other apps (0.9.2)
Switching from Aqua Voice, superwhisper, Wispr Flow, or Handy? Onboarding can pull your old setup over.
Light theme (0.8.1)
A proper light mode. Your eyes might not thank you at first.
Auto Dictionary (0.8.6)
Fix a word during dictation and Glimpse offers to add it to your dictionary automatically.
Local API (0.9.0)
Speech transcription as an OpenAI-compatible endpoint running on your machine.
CLI (0.9.0)
Drive Glimpse from the terminal with the glimpse command: transcribe files, search history, manage models and your dictionary, and script your own integrations.
More
Glimpse-Speech (0.7.0)
Custom local transcription backend, built and tuned for Glimpse. Whisper got faster on device.
Auto-update when idle (0.7.3)
Install updates after the app has been idle a while.
Auto-delete recordings and transcripts (0.9.0)
Schedule cleanup for old audio, transcripts, or both.
Parakeet Unified streaming (0.9.9)
One English model for both fast file transcription and live streaming. About a quarter the size of the old streaming model, faster, and with word-level timestamps.
Personalization snippets (0.8.6)
Modes can inject live context like {{time}} or {{site}}.
Media pause and duck (0.9.2)
Pause playback or lower audio while you dictate, system-wide on macOS and Windows.
Launch at login (0.8.1)
Autostart when you log in. Background launch can open straight to the menu bar (0.9.1).
BYOK cloud transcription (0.9.1)
Transcribe through a cloud speech provider with your own API key when you do not want a local model.
Dictation recovery (0.9.3)
Quit mid-dictation? Glimpse can still save the recording to History when you come back.
Guided onboarding and model browser (0.9.4)
Tell Glimpse your language and whether you want quality, balance, or size, and it picks and downloads the right local model. Browse and filter every model from one place.
Library overhaul (0.9.6)
A cleaner Library with speaker assignment, word-level playback, and better performance.
Mouse buttons as keybinds (0.9.7)
Trigger dictation with a mouse button, not just the keyboard.
Silero VAD (0.9.8)
Higher quality voice detection, so silence stays silent and dictation stays accurate.
Automatic speaker detection (1.0.0)
Library transcripts can label who is talking on their own, with providers that support it.
Homebrew install (1.0.0)
Install the app and the glimpse command together with brew install --cask glimpse-hq/tap/glimpse.
new features
bug fixes
Want to read more? Read the changelog