Description
Hurston-1 is an advanced text-to-speech (TTS) model developed by Destined AI, focusing on accented English speech synthesis. This model is specifically designed to cater to accents often overlooked by conventional TTS systems, such as Southern American, Nigerian, and various African English accents. Built on the Maya1 base model, Hurston-1 introduces Narrowband TTS (NB-TTS), optimized for phone line audio quality, delivering 8 kHz G.711 audio suitable for telephony systems like Twilio.
Read More
The model was trained using a diverse dataset, including real phone call recordings and studio-quality speech, ensuring a realistic and clear output. It supports various voice descriptions, including gender, age, accent, style, and pace, allowing users to generate synthetic voices that sound natural and conversational. Hurston-1 is part of the ClarityVoice family, representing Destined AI's commitment to inclusivity in AI voice technology.
Hurston-1 is ideal for applications such as speech recognition testing, simulated callers for voice agents, training data augmentation, and voice prompts for telephony systems. It also serves educational purposes, providing listening practice for learners and accent familiarity for customer service staff. However, the model does not support consistent voice across sessions or cloning specific individuals.
The model's training data includes contributions from Destined AI's accented English corpus, LibriTTS-R, and Mozilla Common Voice, ensuring a robust and diverse foundation. While Hurston-1 excels in accent accuracy, it is recommended to use the --check option for critical applications to ensure intelligibility and accuracy.
- CategoryAI Models & LLMs
- Pricing ModelOpen source
- DeploymentCloud
- PlatformsWeb
Hurston-1 Highlights
Accented English TTS
Narrowband TTS for telephony
Supports Southern American and Afro-English accents
Built on Maya1 model
8 kHz G.711 audio output
Voice description customization
Part of ClarityVoice family
Open-source model
Getting Started with Hurston-1
Access model: Visit Hugging Face page
Authenticate: Log in or sign up
Set up environment: Follow instructions
Integrate via API: Use provided endpoints
Optimize: Customize voice descriptions
Hurston-1's Use Cases
- Speech Recognition Testing
- Simulated Callers
- Training Data Augmentation
- Voice Prompts
- Education and Training











