Shypd.aiThe largest open index of AI toolsGlow-TTS is an Open Source text-to-speech tool that generates mel-spectrograms from text in parallel. It uses a generative flow via monotonic alignment search, offering fast, diverse, and controllable speech synthesis.
Trending
About
Glow-TTS is an open-source generative flow model designed for text-to-speech (TTS) synthesis, utilizing a monotonic alignment search. Unlike many parallel TTS models, Glow-TTS does not require external aligners, making it a self-contained solution for generating mel-spectrograms from text. By combining the properties of flows and dynamic programming, it efficiently searches for the most probable monotonic alignment between text and the latent representation of speech. This approach ensures robust TTS, capable of generalizing to long utterances, and enables fast, diverse, and controllable speech synthesis. The model achieves significant speed-up over autoregressive models like Tacotron 2 with comparable speech quality and can be extended to multi-speaker settings. It also supports integration with vocoders like HiFi-GAN for improved synthesis quality.
Capabilities
Pricing & Plans
Open Source
Free
FAQs
Trending
