Support
Help and answers for Murmur.
Murmur is system-wide dictation for macOS. A small Swift speech daemon built on Apple's Speech framework runs behind a local loopback API, and an Electron and React desktop app spawns and manages it, so there is no separate service to run by hand. Tap Option anywhere to start talking, and the transcript is pasted into whatever app is frontmost. No account, no cloud, no telemetry. Murmur is pre-release software and currently ships as an Apple Silicon build.
Permissions
- Microphone, required to capture your voice for dictation.
- Speech Recognition, required so Apple's Speech framework can transcribe audio.
- Accessibility, required for the global Option-tap listener and to paste the transcript into the frontmost app.
Requirements
- macOS on Apple Silicon (M-series). The current build is arm64 only.
- Microphone, Speech Recognition, and Accessibility permissions, granted on first run.
- Pre-release software. Expect rough edges and breaking changes between versions.
Privacy
- Audio capture and speech recognition are performed by a local Swift daemon using Apple's Speech and AVAudioEngine frameworks.
- The speech daemon listens only on a loopback address (127.0.0.1); nothing it does is exposed to the network.
- Murmur has no account, no sign-up, no analytics, and no third-party telemetry SDKs.
- Transcripts are pasted straight into the frontmost app and are not stored by Murmur.
- Apple's Speech framework may process recognition requests according to Apple's own privacy policy; Murmur adds no server of its own.
Get in touch
Still stuck? Email and you'll get a real reply from a real person, usually within a day.