Text-to-speech voice trained on the same recordings as Fritter, built on Qwen3-TTS. CC-BY-NC-4.0.