Hurston-1

Hurston-1 is Destined AI’s 3B BF16 Maya1 fine-tune for accented-English text-to-speech, including telephone-quality output for ASR and IVR testing. Hugging Face artifact access is gated; GetLLMs has not downloaded or executed it.

Voice synthesis models
Compare other voice models and documented input/output roles.
F5-TTS Vietnamese
A different language-specific synthesis model.

Voice SynthesisText-to-SpeechTelephonyEnglish Accents

Self-hosted gated Hugging Face artifact; no publisher hardware minimum verified.

Scoped Terms

🚀Function Overview

Hurston-1 generates accented English at 24 kHz or phone-mode 8 kHz G.711, with gated Apache-2.0-labelled weights and no verified hosted API price.

Key Features

3B BF16 fine-tune of maya-research/maya1; publisher labels model weights Apache-2.0.
English only; regional Southern American and African English descriptions, not voice cloning.
Default 24 kHz audio; phone mode applies 300–3400 Hz filtering and outputs 8 kHz G.711 mu-law or A-law.
Voice descriptions include gender, age, accent, style and pace; each call can produce a new voice.
Long text is split around 40-word chunks and voice consistency may change.
Names, dates and amounts remain difficult; the publisher recommends check mode and listening before important use.
Publisher WER/listening/ accent-classifier figures are self-reported, not independent GetLLMs evaluation.
Public model metadata is visible, but weights/code/demo access requires the Hugging Face gate. No hosted provider or API contract was verified.

Use Cases

•Accented speech test data for ASR systems.
•IVR/voice-agent simulated callers and phone prompts.
•Synthetic speech drafts with human listening checks.

⚙️Input Parameters

text

string

English text for local synthesis.

voice description

string

Documented gender, age, accent, style and pace labels.

phone

boolean

Local phone-channel conversion, not a complete real-time phone-agent API.

check

boolean

Optional publisher-described Whisper/speaker checks; not executed here.

💡Usage Examples

Example 1

Input Parameters

{
  "command": "python hurston.py --model . --check --phone --accent \"Southern American accent\" --gender Female --text \"Thanks for calling. How can I help you today?\" --out call.wav"
}

Output Results

Illustrative request or command only. GetLLMs has not executed this example or verified model output.

Quick Actions

Technical Specifications

Hardware Type: Self-hosted gated Hugging Face artifact; no publisher hardware minimum verified.
Commercial Use: The publisher labels model weights Apache-2.0; artifact gating and model/data/demo licenses remain separate.