Enki: Efficient Memory for AI Agents

🧠 Enki — Efficient Memory for AI Agents

The Enki engine uses ~2x less memory for facts (138 vs. 283 for mem0) while maintaining high accuracy and outperforming competitors in multi-session reasoning according to the LongMemEval benchmark.

🌍 Memory optimization reduces context costs when scaling AI agents.

👤 This is a path toward creating faster, cheaper assistants with excellent memory.

Source 1: https://github.com/stephen487/enki-benchmarks