> ## Documentation Index
> Fetch the complete documentation index at: https://help.heymumble.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Install and set up

> Download Mumble Dictation, grant two permissions, download the on-device models, and start dictating in a few minutes.

Setting up Mumble Dictation takes a few minutes: download the app, grant two permissions, let the models download, and pick your shortcut.

## Before you start

Check that your Mac meets the [requirements](/dictation/requirements): Apple Silicon, macOS 14.2 or later, and ideally 16GB of memory or more. You also need an internet connection and about 6GB of free disk space for the one-time model download.

## Set up in five steps

<Steps>
  <Step title="Download and open the app">
    Download the installer from [heymumble.com/dictation](https://heymumble.com/dictation). Open the downloaded `Mumble-Dictation-Mac-latest-arm64.dmg` file and drag Mumble Dictation into Applications, then launch it.
  </Step>

  <Step title="Grant two permissions">
    Mumble asks for exactly two permissions. **Microphone** lets it capture what you say; macOS shows its own permission prompt, so just click **Allow**. **Accessibility** lets Mumble type your words into other apps; macOS does not prompt for this one, so click **Open System Settings** and switch Mumble Dictation on in the Accessibility list.

    <Frame>
      <img src="https://mintcdn.com/naryaai-16a61dd2/58i8Ss9BU7wm9J1-/images/dictation/onboarding-permissions.png?fit=max&auto=format&n=58i8Ss9BU7wm9J1-&q=85&s=b6ad730b22082f33dea415c76e2bb094" alt="The permissions step of setup, asking for Microphone and Accessibility" width="960" height="1120" data-path="images/dictation/onboarding-permissions.png" />
    </Frame>
  </Step>

  <Step title="Download the models">
    Click **Download model**. The voice models (about 2GB) come first, and you can start dictating the moment they finish. The larger language model keeps downloading in the background; you can see its progress inside the app.
  </Step>

  <Step title="Choose language and shortcut">
    Pick the languages you speak, or leave **Auto** on to let Mumble detect each phrase. Then pick your dictation key. Fn is the default; you can choose another key if Fn is taken.
  </Step>

  <Step title="Say something">
    Click into any text field, hold your dictation key, and talk. Release the key and your words land at the cursor.
  </Step>
</Steps>

After setup, Mumble Dictation lives in your menu bar as a small monochrome icon near the clock. Closing the window keeps it running, so your shortcut always works.

## If something goes wrong

* **A permission was denied earlier.** macOS will not ask again. Open **System Settings > Privacy & Security**, then enable Mumble Dictation under **Microphone** and **Accessibility**. Setup cannot continue until both are granted.
* **The model download stops.** Click **Retry**. It picks up where it left off, so a flaky connection never restarts you from zero.
* **Anything else.** See [Troubleshooting](/dictation/troubleshooting) for the full list of fixes.

## Related

* [Requirements](/dictation/requirements)
* [Your first dictation](/dictation/first-dictation)
* [Troubleshooting](/dictation/troubleshooting)