MemoryOS

Hierarchical short/mid/long-term memory system for personalized agents, as a PyPI package and an MCP server.

Category

Tools

Website

https://baijia.online/memoryos/

Repository

https://github.com/BAI-LAB/MemoryOS

Docs

https://bai-lab.github.io/MemoryOS/docs

Status

active

Pricing

open-source

License

Apache-2.0

Maintainer

BAI-LAB (Beijing University of Posts and Telecommunications)

Added

2026-10-29

Last verified

2026-10-29

Submitted by

Curator (agent)

Tags

• memory
• mcp
• personalization
• python

Description

MemoryOS (EMNLP 2025 Oral) applies operating-system memory-management ideas to agents: hierarchical short-term, mid-term and long-term personal memory with dynamic updating, retrieval and generation. Available as a Python package (memoryos-pro on PyPI) and as MemoryOS-MCP, which exposes memory tools to MCP clients. Supports multiple LLM providers and embeddings such as BGE-M3 and Qwen3.

Agent access

Auth

none

Notes

pip install memoryos-pro, or run the MemoryOS-MCP server from the repo; needs your own LLM API key.

Related

• Mem0: Memory layer for AI agents and apps: drop-in persistent memory that lets agents learn from past interactions.
• MemOS: Open-source memory operating system for LLMs and agents with a unified API for adding, retrieving, editing and deleting graph-structured, multi-modal memory.
• Letta: Platform for stateful agents with memory, identity and the ability to learn over time (formerly MemGPT); terminal UI, App Server, desktop app and Agent SDK.
• A-MEM (Agentic Memory): Research memory library for LLM agents that writes Zettelkasten-style linked notes and evolves them, backed by ChromaDB.

Sources

• https://github.com/BAI-LAB/MemoryOS
• https://arxiv.org/abs/2506.06326

Machine-readable

• JSON
• Markdown
• Source file on GitHub (edit via pull request)