Pre-Installed Chatterbox TTS

Overview

Chatterbox TTS is a family of state-of-the-art, open-source text-to-speech and voice cloning models developed by Resemble AI. It provides a comprehensive range of models—from the efficient 350M-parameter Chatterbox-Turbo optimized for low-latency voice agents, to a 500M-parameter Multilingual model supporting 23 languages, and a Single Language Pack with dedicated finetunes for priority languages.

Select a pre-installed Chatterbox TTS Marketplace image and get a fully configured server that's ready to use immediately—no manual installation or complex setup required.

Supported Operating Systems

The Chatterbox TTS Marketplace image is available for Ubuntu Server 22 LTS 64-bit.

Supported Products / Plans

The Chatterbox TTS Marketplace image is currently available across the following plans:

GPU VPS

• Professional GPU VPS - RTX A4000
• Advanced GPU VPS - RTX 5090

GPU Dedicated Server

• Basic Dedicated GPU Server - RTX 5060
• Advanced Dedicated GPU Server - RTX 3060 Ti
• Advanced Dedicated GPU Server - RTX A5000
• Enterprise Dedicated GPU Server - RTX 5090
• Enterprise Dedicated GPU Server - RTX 4090

View compatible plans →

Chatterbox TTS Version

Database Mart offers the latest version of Chatterbox TTS, pre-installed and ready to use.

What's Installed

When you select the Chatterbox TTS Marketplace image, Database Mart automatically installs and configures:

Core Operating Environment

• Operating System: Ubuntu Server 22.04 LTS, pre-optimized for AI workloads
• GPU Drivers & Acceleration: NVIDIA drivers, CUDA 11.8 or 12.1, and cuDNN pre-installed for GPU acceleration
• Python Environment: Python 3.10 with a dedicated virtual environment (Conda or venv) for dependency management

Chatterbox TTS Core Components

• Chatterbox TTS Model: Pre-trained model weights (Chatterbox-Turbo, Chatterbox-Multilingual, or Single Language Pack) downloaded automatically
• Deep Learning Frameworks: PyTorch with CUDA support, Transformers library, and other essential libraries
• Inference Optimizations: MPS support for Apple Silicon, CUDA acceleration for NVIDIA GPUs, and memory-efficient generation

Web Interface & API Server

• Web UI**: Gradio or FastAPI-based web interface for text-to-speech generation with voice cloning capabilities
• REST API Server: Pre-configured OpenAI-compatible API endpoint for easy integration
• Voice Cloning Support: Zero-shot voice cloning from 5-30 seconds of reference audio
• Multilingual Support: Pre-configured for up to 23 languages (en, es, fr, de, ja, zh, etc.)

Hardware & GPU Support

• GPU Memory Management: Resident and offload modes for efficient VRAM usage
• Real-time Generation: Optimized for low-latency voice applications with the Chatterbox-Turbo model

Installation Process

During server provisioning, Database Mart automatically performs the following steps:

1. Deploys the selected Linux operating system (Ubuntu Server 22.04 LTS).
2. Installs the selected Chatterbox TTS model version with all pre-trained weights and configuration files.
3. Installs and configures all required components, including:
   • Python 3.10 with a dedicated virtual environment (Conda or venv)
   • PyTorch with CUDA support for GPU acceleration
   • Chatterbox TTS package and all dependencies
   • Web UI for browser-based speech generation
   • REST API server (OpenAI-compatible endpoint) for easy integration
   • Voice cloning capabilities with reference audio support
   • Multilingual support for up to 23 languages

Everything is fully automated—your Chatterbox TTS server is ready to use as soon as deployment is complete.

Getting Started After Deployment

After your server has been provisioned, follow these steps to begin using Chatterbox TTS.

Step 1: Access the Server

Once deployment is complete, connect to your server via SSH using the credentials provided by Database Mart:

ssh administrator@your-server-ip

The Chatterbox TTS model files are pre-installed in the designated directory (typically /opt/chatterbox or a similar location).

Step 2: Verify the Installation

Check that all components are properly installed and the GPU is ready:

# Verify CUDA and GPU availability
nvidia-smi

# Check Python environment
python3 --version

# Navigate to the Chatterbox directory
cd /opt/chatterbox

# List model files
ls -la

The model weights should be present in the model directory. The Chatterbox-Turbo model is approximately 350M parameters and requires at least 16GB VRAM for optimal performance.

Step 3: Start the Web Interface

The server comes with a pre-configured web UI for text-to-speech generation. Start the service using:

# Activate the virtual environment
cd /opt/chatterbox
source venv/bin/activate

# Start the web interface
python gradio_tts_app.py

or for the FastAPI server:

# Start the API server
python api_server.py

The API server will start and be accessible at http://your-server-ip:8000 with interactive API documentation available at http://your-server-ip:8000/docs.

Step 4: Test the API with a Simple Request

Once the service is running, you can test it using the REST API:

Health Check:

curl -X GET "http://localhost:8000/health"
Expected response:
{
  "status": "healthy",
  "device": "cuda",
  "model_loaded": true
}

Basic Text-to-Speech:

curl -X POST "http://localhost:8000/generate" \
  -H "Content-Type: application/json" \
  -d '{
    "text": "Hello! This is a test of Chatterbox TTS.",
    "exaggeration": 0.5,
    "temperature": 0.7
  }' \
  --output speech.wav

Using Paralinguistic Tags:

curl -X POST "http://localhost:8000/generate" \
  -H "Content-Type: application/json" \
  -d '{
    "text": "Hi there! [chuckle] I am absolutely thrilled to hear from you [excited]!",
    "exaggeration": 0.7
  }' \
  --output speech.wav

Available tags include: [laugh], [chuckle], [sigh], [pause], [excited], [whisper], [cough], and more.

Step 5: Voice Cloning

To clone a voice, provide a reference audio file:

curl -X POST "http://localhost:8000/voice-clone" \
  -F "text=This is voice cloning with a reference audio file." \
  -F "reference_audio=@reference.wav" \
  -F "exaggeration=0.7" \
  --output cloned_speech.wav

The model can clone any voice with just 5-30 seconds of reference audio.

Step 6: Generate Speech with the OpenAI-Compatible API

For OpenAI-compatible integration:

curl -X POST http://localhost:5001/v1/audio/speech \
  -H "Content-Type: application/json" \
  -d '{"input": "Hello, this is a test.", "voice": "alloy"}' \
  --output speech.wav

Step 7: Python Client Example

import requests

# Generate speech
response = requests.post("http://localhost:8000/generate", json={
    "text": "Hello from Python!",
    "exaggeration": 0.5,
    "temperature": 0.7
})

if response.status_code == 200:
    result = response.json()
    audio_id = result['audio_id']
    
    # Download the audio file
    audio_response = requests.get(f"http://localhost:8000/generate/file/{audio_id}")
    with open("output.wav", "wb") as f:
        f.write(audio_response.content)

Notes

Hardware Requirements

Chatterbox TTS models have specific hardware requirements for optimal performance:

Key Features

• Zero-shot Voice Cloning: Clone any voice with just 5-30 seconds of reference audio
• Paralinguistic Tags: Control emotion and expression with tags like [laugh], [excited], [pause]
• Multilingual Support: Supports 23 languages including English, Spanish, French, German, Japanese, and Chinese
• Watermarking: Built-in imperceptible watermarks for responsible AI
• Emotion Exaggeration: Control emotion intensity with the exaggeration parameter (0.0-2.0)

Generation Parameters

Frequently Asked Questions

Q: What are the hardware requirements?

• Minimum: 8GB VRAM for Single Language Pack
• Recommended: 16GB+ VRAM for Chatterbox-Turbo
• Optimal: 24GB+ VRAM for best performance

Q: What is voice cloning and how does it work?

Voice cloning allows you to synthesize speech in any voice using just a reference audio sample. Chatterbox TTS supports zero-shot voice cloning with 5-30 seconds of reference audio. The model captures the speaker's unique characteristics and applies them to the input text.

Q: What languages does Chatterbox TTS support?

Chatterbox TTS supports 23 languages including: Arabic, Danish, German, Greek, English, Spanish, Finnish, French, Hebrew, Hindi, Italian, Japanese, Korean, Malay, Dutch, Norwegian, Polish, Portuguese, Russian, Swedish, Swahili, Turkish, and Chinese.

Q: What are paralinguistic tags?

Paralinguistic tags are special markers in your text that control emotion and expression. Examples include:

• [laugh], [chuckle], [giggle] - Laughter and amusement
• [excited], [happy], [sad] - Emotional states
• [pause], [breath], [whisper] - Vocal effects
• [sigh], [cough], [gasp] - Action sounds

Q: How do I control emotion intensity?

Use the exaggeration parameter (range 0.0-2.0). Higher values produce more expressive and emotional speech; lower values produce more neutral, controlled speech.

Q: What are the common generation parameters?