Helpful only when you ask
Minyma Voice can help with drafts, translation, and local navigation, including weather questions, but message content is not read for voice or translation unless you ask or approve it.
Minyma Voice is separate from your encrypted messaging and calls. Your contact list, your calls, voice notes, photos, videos, files and contact cards are never sent to the speech provider, and starting Voice does not upload your chats. While a Voice session is running, the provider processes your microphone audio, generates the assistant's replies, and receives a small amount of app context so it can act on what you say: your Minyma ID, the names and identifiers of up to ten of your chats, which screen and chat you have open, and your local date, time and language. Nothing from inside a chat is sent on its own. A feature you ask for may send what it needs: a draft you dictate, a poll's question and results, or the text of one message. Private message text is sent only with your permission, for one action or for the current Voice session, one message at a time.
Under the provider's default API policy, content processed by Minyma Voice may be retained in abuse-monitoring logs for up to 30 days, or longer where required by law or reasonably necessary to protect the service or others from harm. Approved retention controls may change this handling. API content is not used to train the provider's models by default.
- Sensitive actions are staged for review
- Current location is requested only for local weather
- Voice notes are encrypted as attachments
- Turning voice off hides product voice controls
- Local pickers do not send media until you choose it