Resemble AI releases DramaBox, open-weight voice model on LTX 2.3
Resemble AI released DramaBox, an open-weight voice synthesis model built on LTX 2.3 that targets expressive text-to-speech generation, with weights, a live demo, and a ComfyUI node now available.
Resemble AI released DramaBox, an open-weight voice synthesis model built on LTX 2.3, on May 13, 2026. The model targets expressive text-to-speech generation and is positioned as capable of rendering emotional tones—anger, surprise, sarcasm—with more prosodic variation than older TTS systems like Tacotron or FastSpeech. Weights are available on HuggingFace under Apache 2.0 license, alongside a live Space demo and GitHub repository.
