by Resemble AI
Open-weights expressive text-to-speech model from Resemble AI with voice cloning, fine-tuned from LTX-2.3 audio.
Updated EU status checked Last shown
Sources: the vendor's own pages (linked above) and creator videos at the linked timestamps. EU status is information, not legal advice; see how we check.
DramaBox is an expressive text-to-speech model from Resemble AI, built as an IC-LoRA fine-tune of the audio-only branch of Lightricks' LTX-2.3 (3.3B parameters, diffusion transformer with flow matching). It is conditioned on Gemma 3 12B text embeddings. The model is English-only.\n\nThe prompt controls the output: dialogue goes inside double quotes and stage directions outside, describing speaker identity, emotion, delivery, laughs, sighs, breaths and pauses. An optional voice reference clip of about 10 seconds (adjustable 3–30 s) clones the target timbre; without one, the model picks a voice matching the description. Parameters include CFG scale, skip-token guidance, duration control and seed, and a watermark option that is on by default.\n\nWeights and code are available on Hugging Face and GitHub, with a ZeroGPU demo Space. The model's inference code uses a warm-server mode of about 2.5 s per generation. Per a YouTube reviewer, it needs a roughly 7 GB audio model plus an 8 GB Gemma text encoder and about 24 GB peak VRAM. It is released under the LTX-2 Community License, which is not an OSI open-source license.
Watch the moment: the exact second top creators show it. Videos load only when you press play, via youtube-nocookie.com.
TTS extracted from LTX 2.3 that clones voices, takes scene direction and produces vocalizations like laughs and sighs. It needs a 7 GB audio model plus an 8 GB Gemma text encoder, with 24 GB peak VRAM.
Closest neighbours by meaning, not just by category.
Model
Expressive text-to-speech model with zero-shot voice cloning, derived from the LTX 2.3 audio model.
App
Open-source, local-first voice studio for voice cloning, text-to-speech, dictation and agent voice output.
Model · ElevenLabs
ElevenLabs' expressive text-to-speech model with inline audio tags, 90+ languages and Professional Voice Clone support.
API · ElevenLabs
AI audio platform and API for text-to-speech, speech-to-text, music, sound effects, voice cloning and voice agents.
Model · Fish Audio
Fish Audio's directable text-to-speech model controlled by plain-language bracketed directions, sound tags and multi-speaker scripts.
Model · Microsoft
Microsoft AI's text-to-speech voice model with expressive output across multiple languages.
