This application is part of an interactive demonstration for an example super language variant, StegaPhone.
In a toyish use case, a news anchor embeds fictitious stock tips by selectively mispronouncing words in a financial news broadcast.
For additional information, see the super language paper.
Generate text-to-speech (TTS) audio from broadcast.
Note rate limiting for the TTS endpoint is enforced. Do not excessively run this function to avoid losing access.