Startup Clairva is providing AI companies with licensed real-world video data that captures how people interact in physical environments, according to a Techub News item citing Tech in Asia. The company is focused on supplying compliant video datasets for training physical AI models. The report describes Clairva’s offering as authorized footage rather than scraped or unspecified material, with the emphasis placed on lawful sourcing and real-world human behavior. Its dataset positioning is tied to model training needs for physical AI systems, where video showing interaction with the surrounding environment can serve as input data. No additional operational details, customers, pricing, or funding information were disclosed in the source brief.