speech-synthesizer
TypeScript icon, indicating that this package has built-in type declarations

1.0.0 • Public • Published

speech-synthesizer

NodeJS library providing local TTS synthesis leveraging VITS (Conditional Variational Autoencoder with Adversarial Learning for End-to-End Text-to-Speech) by using ONNX.

Introduction

This library is based on the work from the folks over at Coqui.ai, specifically the code in this repository is based on their TTS work. This library provides a completly local TTS pipeline that can run in both the browser and server side in NodeJS environments.

Install

npm install --save "speech-synthesizer

Usage

import { SpeechSynthesizer } from "speech-synthesizer";

const speechSynthesizer = await SpeechSynthesizer.create();

// This outputs a Float32Array of single channel 22050hz audio data
const audio = await speechSynthesizer.synthesize("Hello world");

Additional Information

This library is still in it's early stages and some of the code needs to be refactored and improved. The external interface at this point should be stable though and ready for consumption. There needs to be improvements made to the cross platform support and optimizations to speed still however this project is perfectly usable as is and should be sufficient for most use cases.

Package Sidebar

Install

npm i speech-synthesizer

Weekly Downloads

8

Version

1.0.0

License

MIT

Unpacked Size

20.4 kB

Total Files

16

Last publish

Collaborators

  • ianmarmour