MemFit Reduces LLM Agent Latency via LLM-Free Memory Operations

MemFit enables efficient long-term memory for conversational agents by using an append-only store for verbatim turn storage and LLM-free indexing via segment summaries. It employs a multi-path retrieval strategy combining lexical, semantic, and cross-encoder signals to avoid expensive LLM calls during memory construction.