This connects Claude to Deep State Representation graphs, the kind used in robotics for scene understanding and spatial reasoning. You get twelve tools covering the full CRUD cycle: query nodes by type, inspect edges and attributes, insert new nodes and relationships, update attributes, and delete elements. Resources let you read the graph state efficiently without modification. If you're working with robot perception data or need to manipulate spatial scene graphs from an AI assistant, this bridges that gap. The server wraps the Cortex DSR library and requires Python 3.12 plus building Cortex in your virtual environment. Configuration is straightforward via environment variables for agent ID and graph name.