---
title: EsfandTTS
emoji: 🎙️
colorFrom: indigo
colorTo: purple
sdk: gradio
sdk_version: 5.50.0
app_file: app.py
pinned: false
short_description: Qwen3-TTS voice clone (esfand, knut, xqc)
---

# EsfandTTS — Qwen3-TTS voice clone

Voice cloning with [Qwen3-TTS-12Hz-0.6B-Base](https://huggingface.co/Qwen/Qwen3-TTS-12Hz-0.6B-Base).
Cloned voices are built from 12-second reference clips in `voices/` with Whisper
transcripts (`voices/transcripts.json`) for in-context cloning.

## API

```python
from gradio_client import Client

client = Client("trinitytf/EsfandTTS", hf_token="hf_...")  # token needed if private
audio_path = client.predict(
    "Hello chat, this is a test.",  # text
    "esfand",                       # voice: esfand / knut / xqc
    "English",                      # language
    api_name="/predict",
)
print(audio_path)  # local path to the generated wav