Storage Is Not Memory: A Retrieval-Centered Architecture for Agent Recall

TL;DR

The paper introduces True Memory, a retrieval-focused architecture for agent recall that outperforms existing systems on multiple benchmarks by operating over verbatim events without external databases.

Contribution

It presents a novel six-layer retrieval-centered architecture that improves agent memory recall without relying on external storage or complex indexing.

Findings

True Memory Pro achieves 93.0% accuracy on LoCoMo.

It reaches 87.8% on LongMemEval and 76.6% on BEAM-1M.

The architecture outperforms prior methods like Mem0, Supermemory, Zep, and Hindsight.

Abstract

Extraction at ingestion is the wrong primitive for agent memory: content discarded before the query is known cannot be recovered at retrieval time. We propose True Memory, a six-layer architecture that shifts the center of the system from a storage schema to a multi-stage retrieval pipeline operating over events preserved verbatim. The full system runs as a single SQLite file on commodity CPU with no external database, vector index, graph store, or GPU. On LoCoMo (1,540 questions across 10 multi-session conversations), True Memory Pro reaches 93.0% accuracy (3-run mean) against 61.4% for Mem0, 65.4% for Supermemory, approximately 71% for Zep, and 94.5% for EverMemOS under a matched gpt-4.1-mini answer model. On LongMemEval (500 questions), True Memory Pro reaches 87.8% (3-run mean). On BEAM-1M (700 questions at the 1-million-token scale), True Memory Pro reaches 76.6% (3-run mean), above the prior published result of 73.9% for Hindsight. A 56-configuration ablation shows a 1.3-percentage-point spread within the top-performing configuration family.