Expressive text-to-speech model with zero-shot voice cloning, derived from the LTX 2.3 audio model.
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.
Scenema Audio is an audio diffusion model extracted from LTX 2.3. It generates speech from a text prompt with emotional acting, pacing, breath control and sound effects. Prompts use an XML-like <speak> format with a voice description, a scene, an optional language code and an <action> stage direction.
The model separates voice identity from emotional performance. A short, clean reference clip supplies the identity, while emotion comes from the prompt. Per the vendor, a single English reference can produce speech in other supported languages (examples include Spanish, French, German, Italian and Swahili, the last with some phonetic drift). Accent can be steered by the prompt. Voices can also be designed from text alone, without reference audio.
It is offered through the Scenema platform, an AI service for animated long-form video, where Scenema Audio voices in 20+ languages are included in plans. Free, Starter ($8/mo), Creator ($49/mo) and Pro ($209/mo) tiers differ in script length, custom voice clone count (1, 10 or 25 on paid tiers) and other limits. A YouTube reviewer describes an open-source release that runs on about 16 GB VRAM with a quantized model plus 32 GB RAM. No weights or repository link was found in the provided material, so this is unverified.
Watch the moment: the exact second top creators show it. Videos load only when you press play, via youtube-nocookie.com.
Expressive open-source TTS built from the audio model of LTX 2.3. It clones voices, follows stage directions and emotion, changes accents, supports multiple languages and designs voices from text prompts. Needs 16 GB VRAM with a quantized model plus 32 GB RAM.
Closest neighbours by meaning, not just by category.
Model · IndexTeam
Open-weights zero-shot text-to-speech model from Bilibili's IndexTeam with voice cloning and emotion control in 5 languages.
Model · Resemble AI
Open-weights expressive text-to-speech model from Resemble AI with voice cloning, fine-tuned from LTX-2.3 audio.
Model · Edge0
Open-weights 0.17B-parameter multilingual text-to-speech model with zero-shot voice cloning, runs on CPU via ONNX.
Model
Compact 0.6B open text-to-speech model with zero-shot voice cloning and support for 23 languages.
Model · Rednote
Open-source 2B-parameter zero-shot voice-cloning text-to-speech model from Rednote's hilab team.
Model · ByteDance
Research zero-shot TTS model that generates speech by modeling raw waveforms directly with flow-matching DiT.
