On July 16, 2025, Nerval AI launched Palefò's voice/text collection effort with a clear objective: build the first large-scale, production-grade speech platform for Haitian Kreyòl — a language spoken by more than 10 million people worldwide and largely absent from modern AI systems.
The response was historic.
Since launch, contributors across Haiti and the global diaspora have submitted more than 600,000 high-quality voice/text samples, making Palefò one of the fastest language-data mobilizations ever achieved for a low-resource language. What began as a bold experiment has matured into a fully operational AI speech infrastructure.
In early April 2026, Palefò goes live to the public.
A Full-Stack Speech Platform
Palefò is not a research prototype. It is a complete AI speech system built for real-world deployment.
The platform includes:
- A production-ready Text-to-Speech engine delivering clarity, authenticity, and emotional tone unmatched in Kreyòl AI
- A proprietary Automatic Speech Recognition system
- A forthcoming Kreyòl translator application
- A Kreyòl-speaking AI assistant app
Designed for integration across education, media, telecommunications, customer service, digital government, and mobile applications, Palefò enables organizations to deploy high-quality Kreyòl voice capabilities at scale.
Watch the Palefò Intro Video from July 2025
Built to Scale from Day One
Palefò is also the foundation of Nerval's broader language AI architecture.
The linguistic pipelines, dataset frameworks, and model-training systems developed for Kreyòl now power SpeakAI — Nerval's scalable deployment framework for additional high-potential language markets. Rather than building isolated products, Nerval has engineered a repeatable system that can be extended efficiently across dozens of underserved language ecosystems.
Each deployment strengthens the overall platform by increasing data density, improving model performance, and expanding commercial reach.
Infrastructure as Competitive Advantage
To support rapid growth and production deployment, Nerval has built its own AI supercomputing environment powered by NVIDIA GPUs, leveraging its selection into the NVIDIA Inception Program. This infrastructure allows us to:
- Train and fine-tune models internally
- Control performance and cost structures
- Scale inference capacity as adoption grows
- Reduce reliance on external cloud bottlenecks
For speech AI companies, compute access determines speed and competitiveness. Nerval invested early in infrastructure, positioning Palefò to meet rising demand immediately upon launch.
The Opportunity Ahead
More than a billion people globally speak languages that remain minimally represented in mainstream AI systems. Haitian Kreyòl is one of the largest and most commercially viable among them.
Palefò demonstrates that high-performance language AI can be built, scaled, and deployed in these markets — and that demand exists when the technology meets real user needs.
As Palefò goes live in April, it marks more than a product launch. It signals the beginning of a scalable language AI platform that serves as both a commercial engine and a technological foundation for Nerval's next-generation products.
Palefò is proof of execution. And it is only the beginning.