A-MEM
Provides an AI-powered memory system for Claude Code, maintaining persistent and evolving contextual understanding across coding sessions.
概要
This server addresses the challenge of fragmented, non-persistent memories in AI coding tools like Claude Code and Claude Desktop. By creating a Zettlekasten-based Model Context Protocol (MCP) memory system, it enables AI agents to maintain a persistent, evolving understanding of complex codebases across multiple sessions and projects. It automatically discovers relationships and connections as new information is added, fostering deeper insights and seamless continuity for AI-assisted software engineering.
主な機能
- MCP Integration: JSON-RPC 2.0 server compatible with Claude Code and Desktop for seamless tool access.
- 5 GitHub stars
- Prometheus Monitoring: Provides comprehensive observability with metrics for operations, LLM latency, and error rates.
- Advanced Embeddings: Supports real embeddings from services like Sentence-transformers and OpenAI for accurate context.
- Core Memory System: Store, retrieve, and evolve code snippets with AI-generated metadata and vector similarity search.
- Workspace Management: Logically group memories by filesystem path or user-defined names for better organization.
ユースケース
- Building a comprehensive, searchable knowledge base for complex codebases to improve AI understanding.
- Improving continuity and contextual awareness for AI-assisted software engineering workflows within Claude Code.
- Enhancing AI coding agents with persistent, evolving memory across multiple development sessions.