Tier 3 · Agent memory Agent memory · fetched 2026-07-02

Agent Memory Systems Survey

A panoramic survey (2026.03) of 30+ agent memory projects across 15 dimensions - standalone memory frameworks plus agents' built-in memory - including LongMemEval scores. Notably documents the trend of agentic/observational memory overtaking pure vector memory.

View source on GitHub

Key takeaways

01
30+ projects compared across 15 dimensions - ideal checklist fodder
02
Documents the field's shift away from vector-only memory
03
Benchmark scores (LongMemEval) ground the comparisons

Flows built on this research

Memory & Context

Episodic, Semantic, and Procedural Stores

Organize agent memory by kind - what happened, what is true, and how to do things - instead of one undifferentiated pile.

4 steps · 60-90 minutes

Memory & Context

Vector vs Markdown Memory: Make the Decision

Choose your memory backend on evidence: when plain files win, when vectors earn their complexity, and how to test it on your own data.

3 steps · 45-60 minutes

From the archive

Verbatim excerpts mined from our local archive of this repository — the prompts, schemas, and patterns worth stealing.

Insight

March 2026: the 'non-vector era' of agent memory

2026.03 rankings (LongMemEval):
| 1 | Hindsight | TEMPR architecture, 'killed RAG' | 91.4% |
| 2 | Supermemory | LLM compression, 'vector DB is dead' | 99% |
| 3 | Letta | self-editing memory, /remember | - |
| 4 | Mem0 | veteran, graph+vector | 49% |
| 5 | Zep | temporal graph, 'old era' baseline | 71% |

Core trend: in March 2026, vector memory was comprehensively crushed by agentic/observational memory - entering the 'non-vector era'.

[translated] A 30-project survey's headline finding: LLM-native compression and temporal-graph systems roughly doubled the benchmark scores of classic vector RAG memory.

README.md

Memory

Six architecture patterns, from vectors to agentic

Pattern 1 Vector-only: embed -> top-K cosine -> inject (ChromaDB, early Mem0; 'similar but irrelevant' failure mode).
Pattern 2 Graph-augmented: entity extraction -> knowledge graph + vectors (Mem0 Graph, Zep, Cognee ECL: Extract/Cognify/Load).
Pattern 3 Temporal knowledge graph: entities + timestamps, conflict resolution, new facts supersede old (Zep, Hindsight TEMPR: 4 memory networks, retain/recall/reflect).
Pattern 4 LLM-native, no vectors: LLM extracts key info -> structured store (SQLite/files); LLM judges relevance at query time (Supermemory 99% LongMemEval, Google Always-On: IngestAgent/ConsolidationAgent/QueryAgent over pure SQLite).
Pattern 5 Self-editing memory: agent calls memory_write()/memory_read() and decides what to keep (Letta/MemGPT, A-MEM).

[translated] The clearest taxonomy of agent-memory architectures in the vault - five patterns with named exemplars and the failure mode each one fixes.

architecture-patterns.md