Oolel Voices: Open-Source Text-to-Speech for Wolof

We are releasing Oolel-Voices, a text-to-speech model for Wolof and mixed Wolof-French speech.

Try the demo →

What it does

• Trained on natural, spontaneous Wolof speech
• Handles mixed Wolof-French text, reflecting how people actually speak in West Africa
• Accepts a short audio clip as a voice prompt to match a speaking style
• Parameters let you control expressiveness, randomness, and voice fidelity

Try it and access the model

• Hugging Face: huggingface.co/spaces/soynade-research/Oolel-Voices

Questions or feedback: contact@soynade.ai