Generative Models

Generative serving is optional around the Runtime. Dynamic Engagement sub_approach generative still scores through the Runtime; the LLM call is configured on the Dynamic Engagement document (prompt_parameters), not by swapping the scoring endpoint.

Ollama

ecosystem.Ai Ollama  is a container that runs Ollama for local text generation.

Chat

ecosystem.Ai Chat Server  serves chat completions from a prompt. Customer-facing chat is not a replacement for /invocations + /response learning.