Mnemo

0 reviews

Mnemo is long-term memory infrastructure for AI agents and LLM apps. Instead of stuffing raw chat logs into a vector DB and hoping, Mnemo extracts atomic facts on ingest, retrieves them with a hybrid 7-signal pipeline (vector + keyword + knowledge-graph + temporal + fact-aware) fused with Reciprocal Rank Fusion, and returns every result with a citation and score — so your agent can ground answers instead of hallucinating. It's fast and consistent (~150ms warm p50 search, measured on live prod, tight tail) and measured in the open: 80.8% strict on the full LongMemEval-S, every category published — not a cherry-picked number. Model-agnostic (bring your own LLM), US data residency, and you can self-host with Docker or use the hosted API. Free tier, no credit card.