CruxAI has two separate voice features. Read aloud speaks a response to you. Dictation turns your speech into text in the composer. Both are deliberate — neither starts on its own.
Read aloud
Select Read response aloud on any CruxAI reply. Only the text of that response is sent for speech synthesis.
Configure it in Settings -> Speech preferences:
- Voice — Nova, Alloy, Ash, Ballad, Coral, Echo, Fable, Onyx, Sage, or Shimmer.
- Speed — 0.5x to 2x.
- HD voice quality — higher-quality synthesis where available.
Dictation
Select the microphone in the composer to start dictating. Your words land in the message box as text and stay there — nothing is sent until you press Send, so you can read the transcript, fix a misheard word, and add to it before submitting.
Privacy
Dictation only runs after you choose the microphone. Read aloud sends only the text of the response you selected to the speech-processing service, and the generated audio is held in a short-lived private cache so replaying the same response does not re-synthesize it.
Read aloud is for reading an existing response. If you want a spoken back-and-forth instead, use Live Voice Mode.