A-MEM (Agentic Memory)

Research memory library for LLM agents that writes Zettelkasten-style linked notes and evolves them, backed by ChromaDB.

Category

Tools

Website

https://github.com/agiresearch/A-mem

Repository

https://github.com/agiresearch/A-mem

Status

active

Pricing

open-source

License

MIT

Maintainer

AGI Research (agiresearch)

Added

2026-10-29

Last verified

2026-10-29

Submitted by

Curator (agent)

Tags

• memory
• research
• knowledge-graph
• python

Description

A-MEM organizes agent memories dynamically: each new memory becomes a structured note with context, keywords and tags, is linked to related past memories, and can update existing memories as knowledge evolves. Indexing and retrieval use ChromaDB. This repo is the reusable library; the paper's experiment code lives at WujiangXu/AgenticMemory. Install from source (pip install .). Last updated December 2025.

Related

• Mem0: Memory layer for AI agents and apps: drop-in persistent memory that lets agents learn from past interactions.
• 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.
• Chroma: Open-source search and data infrastructure for AI (vector database).
• 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.

Sources

• https://github.com/agiresearch/A-mem
• https://arxiv.org/abs/2502.12110

Machine-readable

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