Getting started

Back to docs
The short version: The first time you launch TongueType, a setup wizard walks you through downloading the Whisper model, granting three macOS permissions, and running a quick test transcription. About a minute, start to finish.

Before you start

TongueType needs:

  • macOS 14 or later on an Apple Silicon Mac (M1, M2, M3, or M4). The Whisper model runs on the Neural Engine, which Intel Macs don’t have.
  • About 500 MB of disk space for the speech model.
  • An internet connection for the one-time model download. After that, TongueType works completely offline.

Step 1: Download the speech model

The wizard’s first step pulls the Whisper model from Hugging Face. It’s a one-time download, around 500 MB, and you’ll see a progress bar while it works. If the download fails, click Retry. The most common cause is a flaky network.

The model lives on your Mac after this. No further network calls are needed for transcription, ever.

Once the model is on disk, TongueType starts loading it onto the Neural Engine in the background as soon as the wizard opens. That load takes a minute or two on most Macs, so it runs while you’re reading the rest of the steps. By the time you reach the end of the wizard, it’s usually already ready.

Step 2: Grant Microphone access

Click Enable. macOS will ask if TongueType can use the microphone. Say yes. This is the obvious one: TongueType can’t transcribe what it can’t hear.

Step 3: Grant Input Monitoring (recommended)

Click Enable. System Settings will open to the Input Monitoring panel. Find TongueType in the list and toggle it on.

Why: TongueType watches for the hotkey globally, even when another app is focused. Input Monitoring is the macOS permission that lets it do that.

This step is optional. If you skip it, the hotkey won’t fire globally and live dictation won’t work, but you can still transcribe audio and video files from the menu bar.

Step 4: Grant Accessibility (recommended)

Click Enable. System Settings will open to the Accessibility panel. Toggle TongueType on, then come back to the wizard and click I’ve enabled it.

This permission is optional. By default, TongueType copies transcriptions to your clipboard, and you press ⌘V to insert them. Granting Accessibility lets TongueType deliver the text into your focused app without you needing to touch the keyboard, which is how the app works as an alternate input method for users who can’t reliably type.

TongueType can be a real accessibility win for users with disabilities such as:

  • Arthritis: joint pain and stiffness in the fingers and hands.
  • Carpal tunnel syndrome: nerve compression causing numbness and weakness.
  • Tendonitis / repetitive strain injury: inflammation from overuse.
  • Trigger finger and Dupuytren’s contracture: fingers catching, locking, or pulling inward.
  • Essential tremor and Parkinson’s disease: involuntary shaking or rigidity that interferes with accurate key presses.
  • Cerebral palsy and multiple sclerosis: muscle-control or coordination issues.
  • Stroke (hemiparesis): weakness or paralysis on one side, often leaving only one hand usable.
  • Spinal cord injury: partial or full paralysis affecting hand and arm function.
  • Amputation or limb difference: missing fingers, hands, or arms that prevent standard two-handed typing.

If you skip this step, TongueType still works fully. The transcription lands on your clipboard and a notice tells you to press ⌘V. Nothing is gated on Accessibility.

Step 5: Try it

The wizard ends on a You’re all set page with a small Try it field already focused. Hold the Right Option key (the default), say something into your mic, then let go. Your words appear in the field within a second or two.

What you see depends on which permissions you granted:

  • With Input Monitoring, hold the hotkey while you speak. Without it, live dictation is unavailable, but you can still transcribe audio and video files from the menu bar.
  • With Accessibility, releasing the key inserts the text directly into the focused field. Without it, the text lands on your clipboard and you press ⌘V to paste it in.

If something looks off, the next page covers the dictation flow in more detail, including how to change the hotkey.

What if I close the wizard early?

No harm done. TongueType keeps running with whatever you’ve set up so far, and a Setup Wizard… entry shows up in the menu bar so you can pick up where you left off. Permissions you revoke later behave the same way: the wizard re-appears so you can fix them.

Where to next

  • Dictation: the hotkey, the grace period, double-tap to latch, and how to cancel a recording in flight.
  • Settings & appearance: pick a microphone, change the language, move the recording overlay around.