Atlanta startup open-sources Hurston-1, a telephone-focused TTS model for Southern and African English
Destined AI, based in Atlanta, has released its Hurston-1 text-to-speech model as open source under the Apache-2.0 license. The model is designed to perform well on American English, especially Southern American and regional African English accents. It was fine-tuned on 136 hours of consented voice data drawn from a larger collection the company has gathered since 2021, and it is optimized for narrowband telephone audio. The company says it has already used the model with customers including Amazon and Salesforce, generating over half a million dollars in revenue. The model does not preserve a consistent voice across prompts or support voice cloning, which limits some uses such as podcast production, and the developer plans a future version with improvements.
This is an AI-generated summary. ShortSingh links to the original source for the complete article.
Discussion (0)
Log in to join the discussion and vote.
Log in