Voices

Generate speech and manage voices

How Voices works

Flow: (1) Library — upload or record a 3–30s reference sample to clone your own voice, or use an admin-featured system voice. (2) Generate — pick a voice, type text, press Synthesize. Three ways to shape the output: Voice clone copies pacing and timbre from the reference audio (re-record slower for slower speech). Text punctuation controls pauses — commas = short, periods = longer, ellipsis (…) = very long. Style prompt (optional, in Generate tab) describes the voice in plain language (e.g. 'calm male voice, slow pace, warm tone') and overrides the selected voice. Daily character limit applies to non-admin users; synthesized audio auto-deletes after 7 days.

Library Generate