Provides a graph-based memory server for AI agents following the Model Context Protocol.
Archiledge serves as a specialized, graph-based memory server designed to power AI agents using the Model Context Protocol (MCP). It stores and manages entities and their relations within a knowledge graph, acting as a central memory bank for agent interactions. While primarily optimized for local development and testing with its embedded graph database options, it exposes a suite of MCP tools, enabling agents to dynamically create, read, search, and delete knowledge graph data, facilitating complex tool-calling workflows.
Key Features
01MCP Tools for creating and deleting entities
02MCP Tools for creating and deleting relations between entities
03MCP Tools for reading the entire graph and searching for nodes
04Flexible persistence options including in-memory, embedded Neo4j, and external Neo4j
05Graph-based Knowledge Graph for storing entities and relations
060 GitHub stars
Use Cases
01Developing and testing AI agents that require a structured memory bank
02Prototyping AI systems that leverage knowledge graphs for context management
03Facilitating tool-calling mechanisms for AI models by providing structured data access
