Ali30223/chatterbox-tts

Chatterbox Turbo TTS App

This package converts the uploaded Colab workflow into a Hugging Face Docker Space.

What this does

• —Uses the same Chatterbox Turbo fork from the Colab notebook: chatterbox-tts @ git+https://github.com/devnen/chatterbox-v2.git@master
• —Uses PyTorch CUDA 12.1 wheels similar to the Colab notebook.
• —Runs a clean Gradio web app on port 7860.
• —Supports optional voice cloning by uploading a clean reference audio clip.
• —Handles long text by chunking and joining audio.

Deployment steps on Hugging Face

1. 1.Create a new Space.
2. 2.Select Docker as the SDK.
3. 3.Upload these files:
4. 4.README.md
5. 5.Dockerfile
6. 6.app.py
7. 7.Use a GPU runtime for realistic speed.
8. 8.CPU can work, but it will be slow.
9. 9.Open the Space after build finishes.

Usage

1. 1.Paste text.
2. 2.Optional: upload a clean 8-15 second voice reference.
3. 3.Click Generate Voice.
4. 4.Play or download the WAV output.

Important notes

• —First run downloads model weights into Hugging Face cache.
• —If you use free CPU hardware, generation may be very slow.
• —For Colab-like speed, use a GPU Space.
• —Reference audio must be longer than 5 seconds.
• —Turbo supports expressive tags such as [laugh], [cough], and [chuckle].

Local Docker test

docker build -t chatterbox-tts-app .
docker run --gpus all -p 7860:7860 chatterbox-tts-app

Then open:

http://localhost:7860