Being developed under the umbrella of the Mozilla Foundation (the same foundation who develops the Firefox browser and Thunderbird mail client), Mozilla TTS is an open source framework whose goal is to simplify the creation of Text to Speech (TTS) voices based on artificial intelligence (AI) techniques. In fact, in recent years the application of machine learning to TTS has lead to unbelievable results, allowing the creation of highly natural, human-sounding voices.
In particular, the framework provides some high performance models for text to speech related tasks, supports fast and efficient model training, and ships with some ready-to-use pre-trained models. In addition to this, it ships with some tools to curate TTS datasets, a demo server for model testing, and some notebooks for extensive model benchmarking.
Whenever possible, references to papers from the scientific literature that describe the implemented neural networks, algorithms and techniques are also included.