memento-mcp
If you are the rightful owner of memento-mcp and would like to certify it and/or have it hosted online, please leave a comment on the right or send an email to henry@mcpreview.com.
Memento MCP is a scalable, high-performance knowledge graph memory system designed for LLMs, offering semantic retrieval, contextual recall, and temporal awareness.
Memento MCP is a knowledge graph memory system that provides long-term ontological memory for LLM clients supporting the model context protocol. It uses Neo4j as its storage backend, enabling both graph storage and vector search capabilities. Memento MCP is designed to be resilient, adaptive, and persistent, offering features like semantic search, temporal awareness, and confidence decay. It integrates seamlessly with platforms like Claude Desktop, Cursor, and GitHub Copilot, providing a unified solution for managing complex knowledge graphs. The system supports advanced metadata, rich entity and relation management, and offers tools for semantic search and temporal graph retrieval. Memento MCP is built to handle large-scale knowledge graphs efficiently, with a focus on performance optimization and simplified architecture.
Features
- Semantic Search: Utilizes vector embeddings for meaning-based retrieval, supporting cross-modal and hybrid search.
- Temporal Awareness: Maintains a complete version history of entities and relations, allowing point-in-time graph retrieval.
- Confidence Decay: Automatically reduces confidence in relations over time, with configurable decay settings.
- Advanced Metadata: Supports rich metadata for entities and relations, including source tracking and custom fields.
- Neo4j Integration: Uses Neo4j for unified graph and vector storage, offering scalability and native graph operations.
Tools
Entity Management
Create, update, delete entities and observe
Relation Management
Create, update, delete relationships
Graph Operations
Read the entire graph, search for nodes, and open nodes
Semantic Search
Semantic-based entity search
Temporal Features
Point-of-time query, history query