dots.tts-mf-2steps

This repository provides a standalone dots.tts artifact for fixed two-step inference. The sampling contract is stored in config.json and selected automatically. Sampling options should be omitted.

Quick start

dots.tts \
  --model-name-or-path dots-studio/dots.tts-mf-2steps \
  --text "Hello, this is a two-step MeanFlow voice cloning example." \
  --prompt-audio /path/to/reference.wav \
  --prompt-text "The exact transcript spoken in the reference audio." \
  --output cloned.wav

from dots_tts.runtime import DotsTtsRuntime

runtime = DotsTtsRuntime.from_pretrained(
    "dots-studio/dots.tts-mf-2steps",
    precision="bfloat16",
)
result = runtime.generate(
    text="Hello, this is a two-step MeanFlow voice cloning example.",
    prompt_audio_path="/path/to/reference.wav",
    prompt_text="The exact transcript spoken in the reference audio.",
)

Passing incompatible sampling values raises an error.

Scope and limitations

This model is intended for fixed two-step inference. Other sampling settings are not claimed. High-fidelity voice cloning must be used only with authorization and consent; do not use it for impersonation, fraud, or disinformation.

---

Citation

@article{dotstts2026,
  title         = {dots.tts Technical Report},
  author        = {dots.tts Team},
  year          = {2026},
  eprint        = {2606.07080},
  archivePrefix = {arXiv},
  primaryClass  = {cs.SD},
}

License

Released under Apache-2.0.