The headline feature is the proprietary . While previous models used standard Tacotron 2 or FastSpeech architectures, the new Wiseguy TTS utilizes a diffusion-based vocoder trained on over 15,000 hours of dialogue from films, radio dramas, and podcasts.
April 19, 2026 Subject: Analysis of the latest “WiseGuy TTS” release (v3/new architecture) Prepared for: AI Voice Technology Monitoring Group wiseguy tts new