Commandment
Voice to text, instantly.
Just press a shortcut and speak. Bring your own OpenAI API key, no subscription required. It needs macOS Accessibility permission to type into other apps.
1.9 MB
v0.1.15 · Requires macOS 15.2
Do I need a subscription?
No. Paste an OpenAI API key into Settings and you pay OpenAI for the seconds you actually dictate, rather than a monthly fee to have it sit idle. The key goes into the macOS Keychain.
Where does my audio go?
Audio is captured at 24 kHz mono and streamed to the OpenAI Realtime API on your own key, transcribed by gpt-4o-mini-transcribe. One temporary WAV sits in the system temp directory and is overwritten on the next take. Nothing else receives it.
Why does it need Accessibility permission?
Typing into another app is the one thing macOS puts behind Accessibility, and that is how transcripts reach your cursor. Skip the prompt and nothing breaks: the transcript goes to your clipboard and you paste it yourself.
What are the shortcuts?
Option+D starts and stops recording. Option+Shift+D opens Settings. Both rebind in Settings.
How do I install it?
Download the DMG from the latest GitHub release, or install from the Homebrew tap. Requires macOS 15.2 or later.
brew install --cask mblode/tap/commandment