NeuralVox

NeuralVox is a project to develop free audiobooks generated by AI. It's currently in the early stages of development.

Projects

While NeuralVox is still under development, we've open-sourced several projects.

OpenPhonemizer

OpenPhonemizer is a permissively-licensed (BSD), open source grapheme to phoneme converter (phonemizer) powered by deep learning. It's based on Deep Phonemizer and works with projects that rely on espeak's phonemizer.

License: BSD-3-Clause

StyleTTS 2 API

A fork of StyleTTS 2 with Python and HTTP APIs. Because it relies on espeak's phonemizer, it's currently GPL licensed.

License: GPLv3

About

Audiobooks have long been recognized as a leading format for consuming literature, however they typically cost thousands of dollars to produce and are often sold for high prices. Moreover, many less popular works have never even been narrated.

NeuralVox aims to solve these issues by harnessing the power of artificial intelligence. By using speech synthesis technology, we can produce natural-sounding audiobooks at a fraction of the cost.

We're still in the early stages of development, but we expect to release some samples soon, so stay posted!