Install VoiceMe, grant the two required permissions, and download the AI models. Takes about 5 minutes on a fast connection.
Click the button below to download the VoiceMe DMG installer. The file is about 30 MB.
Download VoiceMeRequires macOS 14 (Sonoma) or later · Apple Silicon (M1 or newer)
Double-click the downloaded VoiceMe.dmg file. A Finder window will appear showing the VoiceMe app and an Applications folder shortcut.
Drag the VoiceMe icon onto the Applications folder. macOS will copy the app. Once done, eject the DMG (right-click → Eject).
VoiceMe is distributed outside the App Store. macOS will block it the first time you try to open it. This is normal — here is how to allow it.
Step A — Right-click the app and choose Open
Right-click (or Control-click) VoiceMe in the Applications folder → click Open
Step B — Click Open in the confirmation dialog
Click Open. VoiceMe will launch. You will not see this warning again.
Already double-clicked and got "cannot be opened"?
If you double-clicked VoiceMe and macOS showed a blocked message, follow this path instead:
VoiceMe will ask for microphone access the first time you press the Voice → Text hotkey. Click Allow in the prompt, or follow the manual steps below.
If you need to enable it manually:
Open: System Settings › Privacy & Security › Microphone
Toggle VoiceMe on (green). If VoiceMe is not listed, open the app first and press the Voice → Text hotkey once.
Accessibility permission lets VoiceMe detect your global hotkey in any app and paste text at your cursor. macOS will prompt you automatically on first use.
How to enable:
Open: System Settings › Privacy & Security › Accessibility
Toggle VoiceMe on (green). You may need to unlock the pane with your Mac password first. If prompted to quit and reopen VoiceMe, do so.
VoiceMe uses local AI models for transcription and text-to-speech. The models are not bundled with the app — they download once and are stored on your Mac. An internet connection is required for this step only.
How to download the models:
~/.cache/huggingface/hub and only download once.Go to Settings → General → Hotkeys and assign your preferred keyboard shortcuts for Voice → Text and Text → Voice. The app will suggest defaults.
Your Voice → Text hotkey does two things depending on how you press it — no extra settings required.
One key. Two behaviours. Pick whichever feels natural in the moment — you don't have to commit to one.
VoiceMe has two hotkeys — one for voice to text, one for text to voice. Set them in Settings → General → Hotkeys. Choose any key combination that doesn't clash with your other apps. The Escape key cancels a recording that's in progress.
If a word keeps coming out wrong — your name, a product name, a technical term — add it once and VoiceMe corrects it on every transcription. Go to Settings → Dictionary → Word Replacements and tap +. Enter what the model gets wrong on the left, and what you actually want on the right. Matching is case-insensitive and whole-word only, so it won't accidentally change partial matches.
Turn this on and VoiceMe runs a second pass over your transcription before pasting — it fixes punctuation, cleans up sentence structure, and removes repeated words. Enable it in Settings → General → Grammar Correction, then download the model (~300 MB, one time). Expect 1–3 extra seconds per dictation. Turn it off any time to go back to instant paste.
Still stuck? Email support@voicemeapp.com and include your macOS version and a description of the issue.