I built an agent memory system for myself and got 90.8% on LongMemEval
Bitemporal knowledge graph beats vector dumps with 90.8% LongMemEval accuracy.
Memory system for AI agents. #1 on LongMemEval — 96.2% (481/500). Beats every published system including Chronos, Mastra, Supermemory, and Emergence. Built solo in 16 days for $1,000.
World record on LongMemEval beats PwC Chronos, built solo in 16 days.
AI engineers building agent memory and retrieval systems
Chronos · Mastra · Supermemory
Bitemporal knowledge graph beats vector dumps with 90.8% LongMemEval accuracy.
Sleep cycles that actively rewrite memories beat passive vector stores like LangChain memory.
Vector memory for agents beats truncation; Rust core with Python/Node/Go bindings.
Read-time memory consolidation beats Mem0 on benchmark; MCP + TypeScript + SQLite, deploy anywhere.
Fixes agent memory poisoning using Jujutsu VCS—solves contradictory beliefs problem elegantly.
Zero-copy SHM beats Playwright's 2.3s capture loop with 7ms agent vision.