Upload a voice recording and convert it into a different voice. Runs on a free shared GPU — expect 10–30 seconds or more.
Balanced general-purpose voice — good first try.
Save the current voice model, links and pitch to switch back in one tap.