🇭🇹KreyòlAI

English → Haitian Creole · Domain-aware translation for humanitarian and healthcare work

On a 250-sentence held-out human-translated medical test set (CMU NESPOLE corpus), English → Haitian Creole:

Model BLEU chrF TER (lower is better)
Google Translate 30.87 58.44 47.90
Base NLLB-200 30.03 56.46 50.56
KreyolAI Healthcare v3 29.87 56.94 49.44

On a separate human-translated educational test set (MIT-Haiti corpus):

Model BLEU chrF
KreyolAI Education v1 28.63 50.90
Base NLLB-200 25.87 48.28

Honest interpretation: these models are not a benchmark win over the best available systems — they perform on par with Google Translate and the base model, and beat the base model on education content. The real value isn't a higher score, it's a free, offline, domain-specialized translator for a language most providers treat as an afterthought.

Google Translate score obtained via the public translation endpoint, for reference only — not an official Google benchmark. Full details on each model's page: huggingface.co/dondodoai

English Input
Domain

Auto detects from your text

Kreyòl Output

If the translation is wrong, submit the correct version. Every correction helps improve the model.

Example Translations
Examples
Domain
Built on NLLB-200 1.3B with domain-specific LoRA adapters · Healthcare · Humanitarian · Education · General
dondodoai · Purpose-built for humanitarian AI in Haiti