---
license: cc-by-4.0
language:
  - en
pipeline_tag: text-to-speech
tags:
- voxtream
- text-to-speech
---

# Model Card for VoXtream

VoXtream, a fully autoregressive, zero-shot streaming text-to-speech system for real-time use that begins speaking from the first word.

### Key featues

- **Streaming**: Support a full-stream scenario, where the full sentence is not known in advance. The model takes the text stream coming word-by-word as input and outputs an audio stream in 80ms chunks.
- **Speed**: Works **5x** times faster than real-time and achieves **102 ms** first packet latency on GPU.
- **Quality and efficiency**: With only 9k hours of training data, it matches or surpasses the quality and intelligibility of larger models or models trained on large datasets.

### Model Sources 

- **Repository:** [repo](https://github.com/herimor/voxtream) 
- **Paper:** [paper](https://arxiv.org/pdf/2509.15969) 
- **Demo:** [demo](https://herimor.github.io/voxtream) 

## Get started

Clone our [repo](https://github.com/herimor/voxtream) and follow instructions in README file.

### Out-of-Scope Use

Any organization or individual is prohibited from using any technology mentioned in this paper to generate someone's speech without his/her consent, including but not limited to government leaders, political figures, and celebrities. If you do not comply with this item, you could be in violation of copyright laws.

## Training Data

The model was trained on a 9k-hour subset from [Emilia](https://huggingface.co/datasets/amphion/Emilia-Dataset) and [HiFiTTS2](https://huggingface.co/datasets/nvidia/hifitts-2) datasets. For more details please check our paper. 

## Citation 

```
@article{torgashov2025voxtream,
  author    = {Torgashov, Nikita and Henter, Gustav Eje and Skantze, Gabriel},
  title     = {Vo{X}tream: Full-Stream Text-to-Speech with Extremely Low Latency},
  journal   = {arXiv:2509.15969},
  year      = {2025}
}
```