MCP Semantic Graph Memory

Example Use Cases

• 1

  Store and retrieve richly typed semantic knowledge (entities, events, concepts, emotions, reasoning chains) with metadata such as confidence, valence, and evidence.

• 2

  Cognitively-inspired retrieval patterns (spreading activation, weighted relationship traversal, fuzzy matching) to surface relevant context and multi-node associations.

• 3

  Create, persist, and query structured reasoning chains and temporal/causal sequences for explainable analysis and decision support.

Description

This server implements a schema-rich knowledge graph (13 specialized node types) with cognitive dimensions like emotional valence, abstraction level, evidence strength, and reasoning chains, all persisted in Neo4j. It provides schema-compliant node/relationship creation, advanced retrieval (spreading activation, APOC path-finding, temporal and causal traces), and tools for creating and exploring reasoning chains and contextual subgraphs. The system is intended to support agents and pipelines that need structured semantic memory, complex association search, and explainable reasoning traces.