Memory is a powerful MCP Server designed to imbue MCP Clients (like Cursor, Claude, and Windsurf) with the capability to remember user-specific information, such as preferences and behaviors, across various conversations. Leveraging advanced vector search technology, it intelligently retrieves relevant memories based on semantic meaning rather than mere keyword matches. Built entirely on Cloudflare's robust serverless stack, including Workers, D1, Vectorize, Durable Objects, and Workers AI, it provides a scalable, efficient, and cost-effective solution for persistent conversational context, ensuring that interactions are always personalized and informed.
주요 기능
01Semantic search with vector embeddings
02Leverages Cloudflare's serverless ecosystem (Workers, D1, Vectorize, Workers AI, Durable Objects)
03Built-in rate limiting
04Cross-conversation memory for MCP Clients
05Isolated namespaces for user data security
060 GitHub stars
사용 사례
01Storing and retrieving user preferences and behaviors for personalized conversational experiences
02Building knowledge bases with advanced semantic search capabilities
03Tracking and organizing project notes, decisions, and updates
