---
title: breeze-tts-2-demo
canonical_url: "https://www.modelscope.ai/studios/Iceberg-Yang/breeze-tts-2-demo"
md_url: "https://www.modelscope.ai/studios/Iceberg-Yang/breeze-tts-2-demo.md"
repository: Iceberg-Yang/breeze-tts-2-demo
chinese_name: "Breeze TTS 2"
last_updated: 2026-09-01
sdk_type: gradio
sdk_version: 5.49.1
downloads: 0
stars: 1
---

# breeze-tts-2-demo

> breeze-tts-2-demo - An interactive demo by Iceberg-Yang on ModelScope. Bilingual speech generation demo with voice design, voice cloning, voice direction, and automatic Whisper Tiny transcription for reference audio.

Iceberg-Yang/breeze-tts-2-demo is an interactive demo (Studio) on ModelScope. built with gradio 5.49.1. also known as 「Breeze TTS 2」.

- **Repository**: Iceberg-Yang/breeze-tts-2-demo
- **SDK**: gradio
- **SDK version**: 5.49.1
- **Downloads**: 0
- **Stars**: 1
- **Last updated**: 2026-09-01

Source: https://www.modelscope.ai/studios/Iceberg-Yang/breeze-tts-2-demo

---

# Breeze TTS 2

Open-weight bilingual (English/Chinese) text-to-speech model with:
- **Voice Design** — Create a voice from a natural-language description.
- **Voice Clone** — Clone a speaker from reference audio and its exact transcript.
- **Voice Direction** — Clone a reference voice while steering tone, emotion, and pace.

Supports inline vocal events: `(laugh)`, `(sigh)`, `(cough)`, `(clears throat)` in English;
`[笑]`, `[叹气]`, `[咳嗽]`, `[清嗓子]` in Chinese.

## Model

- **Weights**: [BreezeBlue/Breeze-TTS-2](https://modelscope.cn/models/BreezeBlue/Breeze-TTS-2)
- **Code**: [breezeblue-ai/breeze-tts](https://github.com/breezeblue-ai/breeze-tts) (Apache-2.0)
- **License**: Model weights and outputs are for research and non-commercial use only.

## ModelScope migration

This Studio migration keeps Voice Design, Voice Clone, and Voice Direction. Model files are downloaded from ModelScope and cached under `/mnt/workspace`. Reference audio is transcribed on CPU with [openai-mirror/whisper-tiny](https://modelscope.cn/models/openai-mirror/whisper-tiny); the transcript remains editable before generation.
