Back to kords.app

Private voice dictation for Mac

Updated 12 October 2026

Private voice dictation means your speech is turned into text on your own Mac and nothing is kept elsewhere. Kords works that way by default: on-device dictation sends no audio anywhere. Cloud mode is optional, and it processes your audio on a server and then discards it.

What “private” should mean for a dictation app

Dictation hears everything you say while it is on, so it deserves harder questions than most apps. Before you trust one, find out:

What Kords sends, and when

Kords has two ways to turn speech into text. On-device is the default and runs on your Mac. Cloud is optional and trades some privacy for higher accuracy. Here is what happens to each kind of data:

On-device (default)Cloud (optional)
Your audioStays on your MacSent to our server and transcription provider, processed, then discarded
TranscriptsStay on your MacReturned to your Mac. Never stored by us
Text for Voices and translationStays on your MacSent to our server and AI provider, tidied or translated, then discarded
ScreenshotsNever takenNever taken
KeystrokesNever loggedNever logged
Anonymous counts and crash reportsOff unless you opt inOff unless you opt in

Cloud mode and Voices stay off until you switch them on in Settings. Signing in does not switch them on.

What Kords never does

What you keep control of

The permissions Kords asks for, and why

PermissionWhy Kords needs it
MicrophoneTo hear you while you hold your shortcut.
Speech RecognitionTo turn your speech into text on your Mac.
AccessibilityTo listen for your shortcut in any app and to put the text at your cursor.

The full detail is in the privacy policy.

Questions people ask

Does Kords send my audio to the cloud?

Not unless you switch Cloud mode on in Settings. In the default on-device mode your audio is processed on your Mac and nothing is sent.

Does Kords store my dictations?

Your history is saved on your Mac only, and you choose how long it is kept. We do not keep your audio or the words you dictate on our servers.

Who processes my audio in Cloud mode?

Our server is in the EU (Ireland). Audio is transcribed by Groq, with OpenAI as a fallback. It is processed and then discarded. Some of these providers are based in the United States.

Does Kords work without the internet?

Yes, in on-device mode. Cloud mode, Voices and server translation need a connection.

Keep reading