MemoryOS
View on GitHub[EMNLP 2025 Oral] MemoryOS is designed to provide a memory operating system for personalized AI agents.
MemoryOS is a hierarchical memory operating system for personalized AI agents, with short-, mid-, and long-term memory plus storage, update, retrieval, and generation modules. Ships a PyPI package, MCP server, ChromaDB and Playground integrations, and reports SOTA results on the LoCoMo benchmark.
Use Cases
Add long-term memory to AI agentsPersonalized multi-turn chat that remembers usersBuild user profile and knowledge memory automaticallyRetrieve relevant history for context-aware responsesInject memory tools into Claude Desktop, Cursor, or Cline via MCPEvaluate memory systems on LoCoMo benchmarkStore agent memory in ChromaDB vector storeManage short/mid/long-term memory layers
Built With
- Language
- Python
- Frameworks
- ChromaDB · MCP · Docker · vLLM · LlamaFactory · OpenAI API · BGE-M3
Tags
memory · long-term-memory · agent-memory · personalization · rag · mcp · chromadb · vector-database · embeddings · hierarchical-memory · context-management · llm · python · locomo · retrieval · user-profile