Loading...
Discovering amazing AI tools


A text-to-speech model that generates ultra-realistic multi-speaker dialogue in a single forward pass.

A text-to-speech model that generates ultra-realistic multi-speaker dialogue in a single forward pass.
Dia is a text-to-speech (TTS) model designed to synthesize ultra-realistic dialogue in one pass. The project emphasizes efficient generation of conversational speech, enabling the model to produce coherent multi-turn or multi-speaker outputs with natural prosody and timing. Published as an open-source GitHub repository by nari-labs, Dia is intended for use by researchers and developers who need high-quality dialogue synthesis for applications such as conversational agents, media production, and speech research.


