Semantic cache

A semantic cache stores previous model responses and returns a saved answer when a new query is similar enough in meaning to an earlier one, reducing cost and latency for repeated questions.

Semantic cache sits in the Retrieval & memory part of the Agentik {OS} glossary, which defines the words used to build and run AI agent systems.

Also called semantic caching.