#1 Persistent memory for AI coding agents based on real-world benchmarks
Terminal score: 0–1000 raw, weighted across 4 dimensions. Public score: 0–10 normalized (shown in the 30-day stars chart above).