AI systems still lack a critical capability: understanding human taste.

We turn how people engage with multimodal content, across images, video, text, and audio, into structured taste and affinity signals. At the core is a multimodal Taste Graph that models the affinities between people, content, and patterns to infer what someone actually likes, not just what they search for.

Galya is designed as a horizontal layer. Rather than replacing your existing data stack, we sit on top of your data sources and enrich them, so agents and AI-native platforms can make decisions that align with individual taste, optimizing for resonance, not just relevance.

As AI agents increasingly act on behalf of users, they need a way to understand taste across content, context, and behavior. Galya is the infrastructure that lets them understand preference and see the world through their users' eyes.