A
licenseNot graded
qualityB
maintenanceEnables agents to store and retrieve LLM query-response pairs using fuzzy cosine similarity matching with configurable thresholds and cache hit/miss telemetry. This reduces inference cost and latency by reusing semantically similar cached responses.
7
MIT