Semantic caching layer for multi-agent LLM systems. Stop paying your LLM to recompute the same thing twice.
Scoring was not performed