Manages a persistent, indexed knowledge graph, enabling AI agents to store, retrieve, and organize information through entities, relationships, and observations.
Graph Memory provides a robust and persistent memory layer specifically designed for AI agents operating on MCP-compatible platforms. It allows AI models to construct and interact with a knowledge graph, capturing entities, their relationships, and relevant observations. The system features a critical 'main' database that acts as a core index, ensuring the AI always has access to essential information and routing details for other specialized 'contexts'. This modular approach enables AI agents to maintain organized, topic-specific memories while preserving a foundational understanding of their operational landscape.