Documentation
Learn how to setup and use VoxSherpa TTS on your Android device.
Quick Start
Follow these steps to get high-quality offline TTS running in minutes:
Install the App
Download VoxSherpa TTS from the Google Play Store or GitHub Releases.
Download a Model
Go to the Models tab in the app or website. Download the .onnx and tokens.txt for your language.
Import & Play
In the app, click 'Add Model', select your downloaded files, and you're ready to generate speech!
Converting Piper Models for Sherpa-onnx
Many Piper TTS models lack the required tokens.txt file needed for Sherpa-onnx compatibility. You can easily convert them using our automated Google Colab workflow.
Step-by-Step Conversion Guide
- Go to Google Colab and create a new notebook.
- Copy our conversion script from Hugging Face (sample.py).
- Paste the code into your Colab notebook (or ask Colab's Gemini assistant to execute it).
- Run the cell using the Play icon. The script will automatically process your files and output the Sherpa-onnx ready
.onnxand token files.
Have questions or need assistance? Reach out via email at codebysonu95@gmail.com.
Engine Support
VoxSherpa supports multiple high-performance neural engines:
Kokoro-82M
The gold standard for offline quality. Human-like prosody and emotion.
Piper
Extremely fast and efficient. Supports hundreds of voices across many languages.
VITS
Reliable and widely supported engine for diverse linguistic requirements.