Exo-Graph
Externalizes LLM memory into a persistent, searchable knowledge graph, enhancing AI systems with advanced semantic and temporal intelligence.
About
Exo-Graph creates an advanced knowledge graph engine that addresses the ephemeral memory limitations of traditional LLMs. It functions as an 'exocortex' by externalizing LLM memory into a Neo4j database, ensuring knowledge persistence across sessions. The tool captures rich context through a Subject-Relationship-Object triplet structure, enhanced with metadata like summary, confidence, and temporal data. This enables powerful capabilities such as hybrid semantic and graph search, automated entity standardization, intelligent negation handling, and temporal conflict resolution, providing AI systems with a scalable, connected, and evolving memory.
Key Features
- Persistent Knowledge Storage in Neo4j
- Hybrid Semantic and Graph Search Capabilities
- Temporal Intelligence and Conflict Resolution
- Automated Entity and Relationship Standardization
- 0 GitHub stars
- Model Context Protocol (MCP) Integration for AI Assistants
Use Cases
- Building and querying domain-specific knowledge graphs from natural language input
- Tracking and resolving changes in relationships and facts over time
- Providing persistent, searchable memory for AI assistants and LLMs