Onyx-TTS

Onyx-TTS is a high-performance, multilingual text-to-speech system developed by 0N Labs. It's built on ONNX Runtime, delivering fast and efficient speech synthesis with minimal resource requirements.

---

Model Details

• Developed by: 0N Labs
• Model type: Text-to-Speech
• Languages: Multiple languages supported
• License: Apache 2.0
• Model size: ~300MB (quantized: ~80MB)

---

Getting Started

Prerequisites

• Python 3.8 or higher
• pip (Python package manager)

---

Installation

1. Clone the repository:

   git clone https://github.com/0N-Labs/onyx-tts.git
   cd onyx-tts
   

2. Create and activate a virtual environment (recommended):

   python -m venv venv
   source venv/bin/activate    # On Windows: venv\Scripts\activate
   

3. Install the required dependencies:

   pip install -r requirements.txt
   

---

Downloading the Model

Download the model files from the Releases page and place them in your project directory:

• onyx-v1.0.onnx (main model)
• voices-v1.0.bin (voice data)

---

Basic Usage

import soundfile as sf
from onyx_tts import OnyxTTS

# Initialize the TTS engine with the downloaded model files
onyx = OnyxTTS("onyx-v1.0.onnx", "voices-v1.0.bin")

# Generate speech
samples, sample_rate = onyx.create(
    "Hello! This is Onyx TTS by 0N Labs.",
    voice="af_sarah",           # See available voices in VOICES.md
    speed=1.0,                  # Adjust speed (0.5–2.0)
    lang="en-us"
)

# Save the generated audio
sf.write("output.wav", samples, sample_rate)
print("Speech generated successfully!")

---

Running the Web Interface

To launch the demo UI:

python web_interface.py

Then visit: 👉 http://localhost:7860

---

Available Voices

To list all available voices:

from onyx_tts import OnyxTTS

onyx = OnyxTTS("onyx-v1.0.onnx", "voices-v1.0.bin")
print("Available voices:", onyx.get_voices())

See the full list in VOICES.md.

---

Troubleshooting

If you run into issues:

1. Ensure all model files are in the correct location
2. Check that all dependencies are installed
3. Verify Python version is 3.8 or higher
4. Visit the issues page for known bugs and fixes

---

License

• Onyx-TTS: MIT
• Onyx model: Apache 2.0

---

Citation

@software{onyx-tts,
  author = {0N Labs},
  title = {Onyx-TTS: High-performance Text-to-Speech},
  year = {2025},
  publisher = {GitHub},
  journal = {GitHub repository},
  howpublished = {\url{https://github.com/0N-Labs/onyx-tts}}
}