Multi-Memory

概要

Multi-Memory is a robust knowledge graph memory server designed to organize and persist information across various isolated contexts, such as work, personal projects, or specific domains. Built upon the `@modelcontextprotocol/server-memory`, it enhances core functionality with SQLite for reliable storage, comprehensive indexing, and transactional integrity. The server features unique ID-based operations alongside dual identification by name and type, custom JSON properties for rich metadata, and an 'override' mode for efficient record updates. It also introduces the token-efficient TOON format for LLM-optimized responses, making it ideal for contextual memory management in AI applications.

主な機能

• Support for TOON (Token-Oriented Object Notation) for compact, LLM-optimized API responses.
• Dual identification allowing operations by unique ID or a name/type composite key.
• Multi-category support with isolated SQLite databases for distinct memory contexts.
• 0 GitHub stars
• Custom JSON properties on entities, observations, and relations for rich, searchable metadata.
• Override mode to update existing records and observations instead of skipping duplicates.

ユースケース

• Storing and managing contextual knowledge for AI agents or applications like Claude Desktop.
• Organizing project-specific information, personal notes, or domain knowledge in isolated categories.
• Building dynamic knowledge graphs to represent software components, dependencies, and their relationships.