Hurston-1 Open Source Narrowband TTS for Regional American Accents
October 24, 2026
Hurston-1 is an Apache-2.0 licensed narrowband text-to-speech model trained on 1M consented audio recordings from 20,000 contributors. The model specializes in Southern American and Regional African English accents to improve inclusivity in voice interaction.
HOW THIS AFFECTS YOU
●
builderYou can integrate specialized regional dialects into your voice products using this open-source model.
●
designerThis allows you to create more culturally representative and authentic user personas in voice-driven UX.