Elamur Research
☰
home / papers / 2025 / A-MEM: Agentic Memory for LLM Agents

A-MEM: Agentic Memory for LLM Agents

Written for the Research library (claude-opus-5-5) from the paper text, in our own words: the license "arXiv — non-exclusive license to distribute" does not allow redistributing the text itself. Original — arxiv.org/pdf/2502.12110

Summary

LLM agents need long-term memory to reuse past experience. Existing memory modules mostly just store and fetch records. Even graph-based designs such as Mem0 depend on fixed schemas and on read/write points that developers hard-code into the workflow, which makes them hard to adapt to new tasks.

The authors propose A-MEM, a memory system modeled on the Zettelkasten note-taking method, in which the LLM itself organizes what is stored. Each new interaction becomes an atomic note. The LLM enriches it with keywords, tags and a contextual description, and the note also gets a dense embedding. The system then fetches similar older notes and lets the LLM decide which ones to link. A new note can also prompt revisions to the descriptions and attributes of existing notes, which the authors call memory evolution.

A-MEM was tested on the long-term dialogue benchmark LoCoMo with six foundation models and six metrics. The authors report that it beats state-of-the-art baselines and keeps retrieval fast at large scale. The work matters because it gives the agent control over how memory is structured, not just over what it retrieves.

Key points

  • Each note holds the raw content, a timestamp, LLM-generated keywords, tags and a context description, a set of links, and an embedding. The embedding is computed by a text encoder over the concatenated text fields.
  • Adding a note triggers two operations. Link generation retrieves the top-k similar notes and asks the LLM whether to connect them. Memory evolution updates older notes. Linked notes form overlapping "boxes", and retrieving one note also brings in its linked neighbours.
  • In the scaling tests (1,000 to 1,000,000 entries), memory use grew linearly and matched MemoryBank and ReadAgent (1.46 MB to 1464.84 MB). Retrieval time rose from 0.31 µs to 3.70 µs for A-MEM, compared with 0.24 µs to 1.91 µs for MemoryBank and about 43.6 µs to 120,070 µs for ReadAgent.
  • On LoCoMo dialogues, t-SNE plots show tighter clusters than a variant without linking and evolution. A moderate retrieval depth k worked best.
  • Code is available in two repositories: WujiangXu/AgenticMemory (benchmark) and WujiangXu/A-mem-sys (production). The stated limitations are dependence on the quality of the underlying LLM and support for text only.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

Fields: Other

Cited by

137 papers of our library cite it (of the 140673 whose bibliographies we read); 1 of them have a page here, 136 not yet.

Cite

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.48550/arxiv.2502.12110 · arXiv 2502.12110

@misc{xu2025mem,
  title = {{A-MEM: Agentic Memory for LLM Agents}},
  author = {Xu, Wujiang and Liang, Zujie and Mei, Kai and Gao, Hang and Tan, Juntao and Zhang, Yongfeng},
  year = {2025},
  doi = {10.48550/arxiv.2502.12110},
  eprint = {2502.12110},
  archivePrefix = {arXiv},
  url = {https://doi.org/10.48550/arxiv.2502.12110}
}

References

39 references, 10 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Sönke Ahrens. How to Take Smart Notes: One Simple Technique to Boost Writing, Learning and Thinking. Amazon, 2017. Second Edition. find
  2. Anthropic. The claude 3 model family: Opus, sonnet, haiku. Anthropic, Mar 2024. Accessed May 2025. find
  3. Anthropic. Claude 3.5 sonnet model card addendum. Technical report, Anthropic, 2025. Accessed May 2025. find
  4. Akari Asai, Zeqiu Wu, Yizhong Wang, Avirup Sil, and Hannaneh Hajishirzi. Self-rag: Learning to retrieve, generate, and critique through self-reflection. arXiv preprint arXiv:2310.11511, 2023. find our card
  5. Satanjeev Banerjee and Alon Lavie. Meteor: An automatic metric for mt evaluation with improved correlation with human judgments. In Proceedings of the acl workshop on intrinsic and extrinsic evaluation measures for machine translation and/or summarization, pages 65–72, 2005. find our card
  6. Sebastian Borgeaud, Arthur Mensch, Jordan Hoffmann, Trevor Cai, Eliza Rutherford, Katie Millican, George Bm Van Den Driessche, Jean-Baptiste Lespiau, Bogdan Damoc, Aidan Clark, et al. Improving language models by retrieving from trillions of tokens. In International conference on machine learning, p… find
  7. Xiang Deng, Yu Gu, Boyuan Zheng, Shijie Chen, Sam Stevens, Boshi Wang, Huan Sun, and Yu Su. Mind2web: Towards a generalist agent for the web. Advances in Neural Information Processing Systems, 36:28091–28114, 2023. find our card
  8. Khant Dev and Singh Taranjeet. mem0: The memory layer for ai agents. https://github. com/mem0ai/mem0, 2024. find
  9. Darren Edge, Ha Trinh, Newman Cheng, Joshua Bradley, Alex Chao, Apurva Mody, Steven Truitt, and Jonathan Larson. From local to global: A graph rag approach to query-focused summarization. arXiv preprint arXiv:2404.16130, 2024. find
  10. Yunfan Gao, Yun Xiong, Xinyu Gao, Kangxiang Jia, Jinliu Pan, Yuxi Bi, Yi Dai, Jiawei Sun, and Haofen Wang. Retrieval-augmented generation for large language models: A survey. arXiv preprint arXiv:2312.10997, 2023. find our card
  11. Daya Guo, Dejian Yang, Haowei Zhang, Junxiao Song, Ruoyu Zhang, Runxin Xu, Qihao Zhu, Shirong Ma, Peiyi Wang, Xiao Bi, et al. Deepseek-r1: Incentivizing reasoning capability in llms via reinforcement learning. arXiv preprint arXiv:2501.12948, 2025. find our card
  12. I. Ilin. Advanced rag techniques: An illustrated overview, 2023. find
  13. Jihyoung Jang, Minseong Boo, and Hyounghun Kim. Conversation chronicles: Towards diverse temporal and relational dynamics in multi-session conversations. arXiv preprint arXiv:2310.13420, 2023. find
  14. Zhengbao Jiang, Frank F Xu, Luyu Gao, Zhiqing Sun, Qian Liu, Jane Dwivedi-Yu, Yiming Yang, Jamie Callan, and Graham Neubig. Active retrieval augmented generation. arXiv preprint arXiv:2305.06983, 2023. find
  15. David Kadavy. Digital Zettelkasten: Principles, Methods, & Examples. Google Books, May 2021. find
  16. Jiho Kim, Woosog Chay, Hyeonji Hwang, Daeun Kyung, Hyunseung Chung, Eunbyeol Cho, Yohan Jo, and Edward Choi. Dialsim: A real-time simulator for evaluating long-term multi-party dialogue understanding of conversational agents. arXiv preprint arXiv:2406.13144, 2024. — [arXiv:2406.13144](https://arxiv.org/abs/2406.13144) find
  17. Kuang-Huei Lee, Xinyun Chen, Hiroki Furuta, John Canny, and Ian Fischer. A human-inspired reading agent with gist memory of very long contexts. arXiv preprint arXiv:2402.09727, 2024. find
  18. Patrick Lewis, Ethan Perez, Aleksandra Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Küttler, Mike Lewis, Wen-tau Yih, Tim Rocktäschel, et al. Retrieval-augmented generation for knowledge-intensive nlp tasks. Advances in Neural Information Processing Systems, 33:9459–9474, 2020. find
  19. Chin-Yew Lin. Rouge: A package for automatic evaluation of summaries. In Text summarization branches out, pages 74–81, 2004. find
  20. Xi Victoria Lin, Xilun Chen, Mingda Chen, Weijia Shi, Maria Lomeli, Rich James, Pedro Rodriguez, Jacob Kahn, Gergely Szilvasy, Mike Lewis, et al. Ra-dit: Retrieval-augmented dual instruction tuning. arXiv preprint arXiv:2310.01352, 2023. find our card
  21. Zhiwei Liu, Weiran Yao, Jianguo Zhang, Liangwei Yang, Zuxin Liu, Juntao Tan, Prafulla K Choubey, Tian Lan, Jason Wu, Huan Wang, et al. Agentlite: A lightweight library for building and advancing task-oriented llm agent system. arXiv preprint arXiv:2402.15538, 2024. — [arXiv:2402.15538](https://arxiv.org/abs/2402.15538) find
  22. Adyasha Maharana, Dong-Ho Lee, Sergey Tulyakov, Mohit Bansal, Francesco Barbieri, and Yuwei Fang. Evaluating very long-term conversational memory of llm agents. arXiv preprint arXiv:2402.17753, 2024. find
  23. Kai Mei, Zelong Li, Shuyuan Xu, Ruosong Ye, Yingqiang Ge, and Yongfeng Zhang. Aios: Llm agent operating system. arXiv e-prints, pp. arXiv–2403, 2024. find
  24. Ali Modarressi, Ayyoob Imani, Mohsen Fayyaz, and Hinrich Schütze. Ret-llm: Towards a general read-write memory for large language models. arXiv preprint arXiv:2305.14322, 2023. — [arXiv:2305.14322](https://arxiv.org/abs/2305.14322) find
  25. Charles Packer, Sarah Wooders, Kevin Lin, Vivian Fang, Shishir G Patil, Ion Stoica, and Joseph E Gonzalez. Memgpt: Towards llms as operating systems. arXiv preprint arXiv:2310.08560, 2023. find
  26. Kishore Papineni, Salim Roukos, Todd Ward, and Wei-Jing Zhu. Bleu: a method for automatic evaluation of machine translation. In Proceedings of the 40th annual meeting of the Association for Computational Linguistics, pages 311–318, 2002. find our card
  27. Nils Reimers and Iryna Gurevych. Sentence-bert: Sentence embeddings using siamese bert-networks. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing. Association for Computational Linguistics, 11 2019. find
  28. Aymeric Roucher, Albert Villanova del Moral, Thomas Wolf, Leandro von Werra, and Erik Kaunismäki. ‘smolagents‘: a smol library to build great agentic systems. https://github. com/huggingface/smolagents, 2025. find
  29. Zhihong Shao, Yeyun Gong, Yelong Shen, Minlie Huang, Nan Duan, and Weizhu Chen. Enhancing retrieval-augmented large language models with iterative retrieval-generation synergy. arXiv preprint arXiv:2305.15294, 2023. find our card
  30. Zeru Shi, Kai Mei, Mingyu Jin, Yongye Su, Chaoji Zuo, Wenyue Hua, Wujiang Xu, Yujie Ren, Zirui Liu, Mengnan Du, et al. From commands to prompts: Llm-based semantic file system for aios. arXiv preprint arXiv:2410.11843, 2024. — [arXiv:2410.11843](https://arxiv.org/abs/2410.11843) find
  31. Harsh Trivedi, Niranjan Balasubramanian, Tushar Khot, and Ashish Sabharwal. Interleaving retrieval with chain-of-thought reasoning for knowledge-intensive multi-step questions. arXiv preprint arXiv:2212.10509, 2022. find our card
  32. Bing Wang, Xinnian Liang, Jian Yang, Hui Huang, Shuangzhi Wu, Peihao Wu, Lu Lu, Zejun Ma, and Zhoujun Li. Enhancing large language model with self-controlled memory framework. arXiv preprint arXiv:2304.13343, 2023. — [arXiv:2304.13343](https://arxiv.org/abs/2304.13343) find
  33. Xingyao Wang, Boxuan Li, Yufan Song, Frank F Xu, Xiangru Tang, Mingchen Zhuge, Jiayi Pan, Yueqi Song, Bowen Li, Jaskirat Singh, et al. Openhands: An open platform for ai software developers as generalist agents. arXiv preprint arXiv:2407.16741, 2024. find our card
  34. Zhiruo Wang, Jun Araki, Zhengbao Jiang, Md Rizwan Parvez, and Graham Neubig. Learning to filter context for retrieval-augmented generation. arXiv preprint arXiv:2311.08377, 2023. find
  35. Lilian Weng. Llm-powered autonomous agents. lilianweng.github.io, Jun 2023. find
  36. J Xu. Beyond goldfish memory: Long-term open-domain conversation. arXiv preprint arXiv:2107.07567, 2021. find
  37. Wenhao Yu, Hongming Zhang, Xiaoman Pan, Kaixin Ma, Hongwei Wang, and Dong Yu. Chain-of-note: Enhancing robustness in retrieval-augmented language models. arXiv preprint arXiv:2311.09210, 2023. find
  38. Zichun Yu, Chenyan Xiong, Shi Yu, and Zhiyuan Liu. Augmentation-adapted retriever improves generalization of language models as generic plug-in. arXiv preprint arXiv:2305.17331, 2023. find
  39. Wanjun Zhong, Lianghong Guo, Qiqi Gao, He Ye, and Yanlin Wang. Memorybank: Enhancing large language models with long-term memory. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 38, pages 19724–19731, 2024. Contents Introduction Related Work 2.1 Memory for LLM Agents . . . .… find