WellSaid Labs Introduces 'HINTS': Revolutionizing AI Voice Customization Standards

WellSaid Labs, a leader in artificial intelligence (AI) voice technology, has launched an innovative tool that permits users to direct AI voice performances with greater naturalness and nuance. This new feature, named HINTS (Highly Intuitive Naturally Tailored Speech), empowers content creators to customize AI voices using contextual annotations, such as tempo and loudness adjustments, similar to the way a movie director works.

Michael Petrochuk, co-founder and CTO of WellSaid Labs, shared in an exclusive interview, “Our customers have expressed a desire for greater control over the vocal outputs of our AI. We aimed to create a system that is both intuitive and natural, enabling our model to predict authentic performances based on user context, so creatives can realize their artistic vision.”

HINTS marks a departure from traditional methods that rely on rigid markup languages or basic prompts for controlling AI voices. This new technology allows for detailed, interpolable adjustments—such as modifying a specific passage to be 0.7x slower or increasing volume by 5 dB—while the AI voice responds seamlessly. Its contextual awareness enables users to layer and nest annotations across extensive scripts.

“The system uses actual human data (consensually obtained) for its audio outputs, making its annotated verbalizations as realistic as those without annotations,” Petrochuk explained. “Remarkably, we found that the model not only effectively utilizes a single dataset but can also generalize across performances from multiple speakers to enhance its prosody. This discovery exceeded our expectations and highlights the potential for future research.”

HINTS meets the demand for highly customizable, director-focused AI voice tools, potentially transforming voice-based content for audiobooks, training modules, marketing videos, and more. Initial evaluations indicate improvements in accuracy and naturalness.

The research also prioritizes responsible and ethical AI practices. “From the beginning, we’ve been committed to ethical innovation,” Petrochuk noted. WellSaid ensures explicit consent from voice contributors, safeguards privacy, and moderates content to prevent misuse.

As vocal AI becomes increasingly integrated into consumer technology and entertainment, HINTS exemplifies how this technology can serve as an empathetic storytelling medium rather than merely a vocal tool. While there are still limitations when compared to human talent, innovations like HINTS bring us closer to achieving truly expressive synthetic voices.