Nocturne Memory

About

Nocturne Memory serves as an external hippocampus for AI agents, providing them with a robust, structured, and long-term memory solution. Built on a Neo4j graph database, it manages knowledge as interconnected entities and relationships, enabling AI to retain conversational history, user preferences, and complex world settings beyond transient contexts. The system facilitates seamless human-AI collaboration, allowing AI to read and write to the knowledge base via the Model Context Protocol (MCP), while humans maintain oversight through a visual interface that supports auditing, diff-viewing, and rollback of AI-generated content. This prevents '7-second memory' for AI and offers a powerful alternative for managing intricate relational data.

Key Features

• 16 GitHub stars
• Neo4j-based dynamic knowledge graph backend with version control
• Model Context Protocol (MCP) server for AI agent read/write access
• Human-in-the-Loop review and audit interface with diff comparison and rollback
• Visual memory explorer for intuitive management of entities, relationships, and chapters
• Automated cleanup for orphan nodes and historical data

Use Cases

• Managing complex world-building for novels, games, or TRPG campaigns
• Enhancing relational note-taking for scenarios requiring advanced graph-based knowledge management
• Providing AI assistants with persistent, structured long-term memory and knowledge