First Run
The first time you open Murmur, it walks you through a short setup flow. This takes about two minutes and gets you ready to start dictating.
Welcome screen
The welcome screen gives a brief overview of what Murmur does: system-wide voice-to-text via global hotkeys, on-device transcription, and optional AI cleanup. Click Get Started to move through the setup steps.
Granting permissions
Murmur needs a few system permissions to function. The onboarding screen explains each one and opens the relevant System Settings panel directly — you do not need to navigate there yourself.
Step 1 — Microphone
Murmur needs microphone access to capture your voice. macOS will show a standard permission prompt the first time. Click Allow.
This is the only permission that is strictly required. Nothing works without it.
Step 2 — Input Monitoring
Input Monitoring lets Murmur listen for your global hotkey from any app — even when Murmur itself is not in focus. Without this, the hotkey only works when the Murmur window is frontmost.
macOS will direct you to System Settings > Privacy & Security > Input Monitoring. Toggle Murmur on in that list.
Step 3 — Accessibility
Accessibility access lets Murmur automatically paste transcribed text at your cursor. Without it, Murmur copies text to the clipboard instead, and you paste manually.
macOS will direct you to System Settings > Privacy & Security > Accessibility. Toggle Murmur on.
Step 4 — Screen Recording (optional)
Screen Recording lets Murmur read your active window title. This is used by the optional AI refinement feature to tailor corrections to your current context — for example, knowing you are in a code editor versus a chat app. If you do not use AI refinement, you can skip this.
You can skip any permission step and grant it later from Settings > Permissions inside Murmur.
Choosing a transcription model
After permissions, Murmur asks you to pick a transcription model. If you are unsure, the recommended choice is Parakeet V2.
- Parakeet V2 — Best accuracy and speed for English. Recommended for most users.
- WhisperKit (small or medium) — Good option if you need language support beyond English.
- Apple Speech — No download required. Uses the system speech recognizer built into macOS Tahoe.
After you select a model, Murmur downloads it in the background. Download size ranges from a few hundred MB to about 1.5 GB depending on your choice. You can start using Murmur while the download finishes, but transcription will not be available until the model is ready.
You can switch models or download additional ones later from Settings > Dictation Models.
Optional setup
AI provider
Murmur can optionally clean up transcriptions using an AI model. This is off by default. If you want to enable it, the setup flow lets you enter an OpenAI API key or configure a local Ollama instance. You can also set this up later in Settings > AI Refinement.
Launch at login
You can opt in to having Murmur start automatically when you log in. Since Murmur lives in the menu bar, most users find this useful. You can change this preference at any time in Settings > General.
You are ready
Once setup is complete, Murmur appears in your menu bar. Press the hotkey you chose during setup to start dictating in any app — there is no preset default, so the shortcut is always one you picked. The next article covers permissions in more detail if you want to understand what each one does and how to adjust them.
Skipped something? The setup checklist lives on the Home page until every step is done, and you can bring it back any time with Reset Onboarding in Settings > Developer.