Personalization API

What it does

The Galya API computes structured taste context and returns preference signals your application can act on: affinity scores for ranking, reranking for generated results, and taste context you can inject into agents. You call it at inference, so personalization reflects the current request, not a stale profile.

Built for developers

Integration is through the Galya SDK and agent packages, with agent tools for passing preference-aware context into autonomous workflows. You do not build or maintain a recommendation stack, and you do not move your data. The API fits alongside your existing model and retrieval layers.

No cold-start, no PII

Because Galya models taste structurally rather than from accumulated behavior, the API returns useful signal on the first interaction, not after a profile has built up. It computes preference signals without personal identifiers, so it stays compliant with GDPR and CCPA by design.

What you can build

Preference-correct ranking and search, taste-based reranking of AI-generated content, and agents that carry a sense of the user's taste into every step. One API, read from the same Taste Graph.

Start building

Add taste and preference signals to your product, and see the first Taste Graph compose against your own use case.