🧩 Caura: Shared Memory Layer for Agent Fleets

Caura (formerly MemClaw) focuses on memory management within multi-agent systems. It provides a governed shared memory layer for fleets of AI agents. The technical direction is interesting: it is MCP-native and built for multi-tenant environments. It introduces trust tiers, audit trails, knowledge graphs, and self-improving retrieval mechanisms. As agent systems scale from single instances to collaborative fleets, securely and efficiently sharing context becomes a core challenge, which Caura attempts to solve at the infrastructure level.