Speech to text on a Mac turns spoken words into text in almost any app, from notes and messages to full documents. This overview explains how the built-in dictation features work, how they compare with third‑party tools, and how to get reliable results.
Whether you want hands‑free writing, accessibility support, or faster note‑taking, macOS includes several options for converting speech to text with strong privacy and integration.
| Feature | Built‑in Dictation | Enhanced Dictation | Third‑Party Apps |
|---|---|---|---|
| Offline use | Yes, with language pack | Yes, requires download | Depends on app and plan |
| Continuous dictation | Limited, tap to start/stop | Yes, up to several minutes | Often full‑session support |
| Commands & shortcuts | Basic punctuation and navigation | Same, more consistent | Custom phrase lists and advanced actions |
| Privacy | Processing on device when possible | On device, no audio stored | Check vendor policies; some require cloud |
| Language coverage | Many languages, varies by version | Same languages, larger models | Often broader, niche languages supported |
Enable Enhanced Dictation on your Mac
Enhanced Dictation keeps audio processing on your Mac, which reduces latency and protects privacy. You can turn it on in System Settings and download the required language model for offline use.
With Enhanced Dictation active, you can speak longer passages without tapping the keyboard repeatedly. This mode is useful for drafting emails, notes, and longer documents in a single session.
Dictation Shortcuts and Activation
Quick activation depends on your Mac model and whether a dictation key is available. On many keyboards, you press the Dictation key once to start speaking and say “stop” to finish.
- Press the Dictation key or use a custom shortcut in Settings.
- Use standard punctuation commands like “new line” or “comma”.
- Choose a quiet environment and speak clearly for best accuracy.
- Train your Mac to recognize your voice by reading sample phrases in Settings.
Accuracy, Noise, and Language Settings
Background noise and microphone quality directly affect accuracy. Use a headset or improve room acoustics to reduce misunderstandings. macOS adapts over time as it learns your voice patterns.
Language and region settings determine which dictionaries and models are available. Selecting the correct language in System Settings ensures proper punctuation, capitalization, and domain‑specific vocabulary support.
Dictation vs Voice Control
Dictation converts speech to text, while Voice Control lets you navigate and click using spoken commands. You can use both together to edit and control your Mac entirely by voice.
Voice Control relies on the same language files as Enhanced Dictation. Keeping both features enabled gives you flexibility for writing and full‑system accessibility without third‑party tools.
Key Takeaways for Speech to Text on a Mac
- Enable Enhanced Dictation for offline, low‑latency transcription.
- Use clear microphones and quiet spaces to improve accuracy.
- Learn punctuation and navigation commands to speed up editing.
- Combine Dictation and Voice Control for full voice workflows.
- Check System Settings regularly for language updates and model refreshes.
FAQ
Reader questions
Why does my speech to text insert wrong words on a Mac?
Background noise, unclear pronunciation, or an untrained voice profile can cause errors. Try improving your microphone, reading the training phrases, and using Enhanced Dictation to reduce mistakes.
Can I use speech to text offline on a Mac?
Yes, with Enhanced Dictation enabled and the appropriate language model downloaded, your Mac can process speech offline without sending audio to the cloud.
How do I add custom vocabulary or shortcuts for frequent phrases?
macOS does not include a built‑in phrase shortcut list, but you can use Text Replacement in Settings or create custom macros in third‑party apps to automate frequently used terms.
Is my dictation audio stored when I use Enhanced Dictation?
No, with Enhanced Dictation your audio stays on your Mac and is not uploaded or stored by Apple after the language model is downloaded.