Backed by

Multimodal data that
reflects the real world.

We collect high-quality multimodal human data, starting with speech and video across diverse expressions, emotions, and interactions, to help you build AI models that engage more naturally with the world.