Building Sulci, the semantic cache for LLM apps and Agentic AI. When a new query means the same thing as a past one, you get the cached answer instead of another model call.
Context-aware: pass a session and the same follow-up in two conversations searches with two different vectors.