Diktator!
Privacy
Diktator! does not collect personal data.
Your dictated audio and transcribed text are processed locally on your Mac and are not sent to our servers. Diktator! does not use analytics, advertising trackers, user accounts, or cloud transcription. The app may connect to the internet only when you explicitly download the local AI model required for speech recognition.
Diktator! does not collect dictated audio, transcripts, dictated text, documents, or personal content. If you send dictated text through Mail, Messages, WhatsApp, a browser, or another app, that sending is controlled by that app or service, not by Diktator!.
macOS asks before Diktator! can use the microphone. Recording starts only with a hotkey or app command, and the capsule shows a green wave for normal dictation or a yellow wave for English output.
The Mac App Store and TestFlight builds do not request or use Accessibility or Input Monitoring. They return text through the standard macOS Services flow where supported, with copy and drag fallback when needed. Diktator+ direct edition may use Accessibility only to place locally recognized text into the active app; it does not use that permission to read documents, inspect apps, monitor activity, or collect typed information.
Diktator! uses local checks around speech recognition: short silence is ignored, temporary audio is cleaned, spoken formatting is handled separately from language cleanup, and final proofreading is guarded so it does not replace good multilingual text with a worse result.
The model download request is only for downloading the model file after you approve it. Dictated audio, transcripts, documents, and personal content are not sent with that request.
Model files stay on your Mac. Temporary audio files are removed after dictation, and diagnostic or temporary files are cleaned automatically. Diktator! does not use analytics, ads, tracking, or user accounts.