I built an agent memory system for myself and got 90.8% on LongMemEval

Name: I built an agent memory system for myself and got 90.8% on LongMemEval
Availability: InStock
Author: inshane

by inshane·Apr 14, 2026·3 points·0 comments

AI Analysis

●●●BangerBig BrainWizardry

Bitemporal knowledge graph beats vector dumps with 90.8% LongMemEval accuracy.

Strengths

•Bitemporal tracking (fact validity vs. learning time) solves temporal reasoning gaps
•Entity resolution and contradiction detection built into the knowledge graph layer
•Works with any MCP client and LLM backend without vendor lock-in

Weaknesses

AI/ML●●Solid

Sleep cycles that actively rewrite memories beat passive vector stores like LangChain memory.

Big BrainNiche Gem

artificium

402mo ago

Security●●Solid

STIX ontology and threat actor alias resolution beat Mem0 and Graphiti for CTI.

Niche GemSolve My Problem

rolandpg

311mo ago

SaaS●Mid

Knowledge graph compression (3,714x token ratio) is impressive, but 'persistent agent memory' is crowded territory.

Solve My Problem

bikidev

103mo ago

AI/ML●●Solid

Temporal knowledge graph memory and trace-to-test evals beat standard vector RAG.

Solve My ProblemBig Brain

baranozdemir

402d ago

AI/ML●●Solid

Unifies siloed agent memory via MCP, but Mem0 and Zep already exist.

Niche GemSolve My ProblemShip It

feelingsonice

512mo ago

AI/ML●●Solid

MESI coherence for agent memory is clever, but LangGraph and AutoGen already own this space.

Big BrainBold Bet

doiito

1016d ago