[{"data":1,"prerenderedAt":4},["ShallowReactive",2],{"readme:lightmem":3},"\u003Cdiv align=\"center\">\n  \u003Cpicture>\n    \u003Csource srcset=\".\u002Ffigs\u002Flightmem_logo_dark.png\" media=\"(prefers-color-scheme: dark)\">\u003C\u002Fsource>\n    \u003Cimg src=\"https:\u002F\u002Fraw.githubusercontent.com\u002Fzjunlp\u002Flightmem\u002FHEAD\u002Ffigs\u002Flightmem_logo_light.png\" width=\"60%\" height=\"40%\" \u002F>\n  \u003C\u002Fpicture>\n\u003C\u002Fdiv>\n\u003Ch1> LightMem: Lightweight and Efficient Memory-Augmented Generation \u003C\u002Fh1>\u003Cp align=\"center\">\n  \u003Ca href=\"https:\u002F\u002Farxiv.org\u002Fabs\u002F2510.18866\" rel=\"nofollow ugc noopener\">\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FarXiv-Paper-red\" alt=\"arXiv\" \u002F>\n  \u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fzjunlp\u002FLightMem\" rel=\"nofollow ugc noopener\">\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fzjunlp\u002FLightMem?style=social\" alt=\"GitHub Stars\" \u002F>\n  \u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fzjunlp\u002FLightMem\u002Fblob\u002Fmain\u002FLICENSE\" rel=\"nofollow ugc noopener\">\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FLicense-MIT-green.svg\" alt=\"License: MIT\" \u002F>\n  \u003C\u002Fa>\n  \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Flast-commit\u002Fzjunlp\u002FLightMem?color=blue\" alt=\"Last Commit\" \u002F>\n  \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FPRs-Welcome-red\" alt=\"PRs Welcome\" \u002F>\n\u003C\u002Fp>\u003Ch5> ⭐ If you like our project, please give us a star on GitHub for the latest updates!\u003C\u002Fh5>\u003Chr \u002F>\n\u003Cp>\u003Cstrong>LightMem\u003C\u002Fstrong> is a lightweight and efficient memory management framework designed for Large Language Models and AI Agents. It provides a simple yet powerful memory storage, retrieval, and update mechanism to help you quickly build intelligent applications with long-term memory capabilities.\u003C\u002Fp>\n\u003Cul>\n\u003Cli>\u003Cp>🚀 \u003Cstrong>Lightweight & Efficient\u003C\u002Fstrong>\n\u003Cbr \u002F> Minimalist design with minimal resource consumption and fast response times\u003C\u002Fp>\n\u003C\u002Fli>\n\u003Cli>\u003Cp>🎯 \u003Cstrong>Easy to Use\u003C\u002Fstrong>\n\u003Cbr \u002F> Simple API design - integrate into your application with just a few lines of code\u003C\u002Fp>\n\u003C\u002Fli>\n\u003Cli>\u003Cp>🔌 \u003Cstrong>Flexible & Extensible\u003C\u002Fstrong>\n\u003Cbr \u002F> Modular architecture supporting custom storage engines and retrieval strategies\u003C\u002Fp>\n\u003C\u002Fli>\n\u003Cli>\u003Cp>🌐 \u003Cstrong>Broad Compatibility\u003C\u002Fstrong>\n\u003Cbr \u002F> Support for cloud APIs (OpenAI, DeepSeek) and local models (Ollama, vLLM, etc.)\u003C\u002Fp>\n\u003Cdiv align=\"center\">\u003Cimg src=\"https:\u002F\u002Fraw.githubusercontent.com\u002Fzjunlp\u002Flightmem\u002FHEAD\u002Ffigs\u002Farchitecture.jpg\" width=\"100%\" height=\"60%\" \u002F>\u003C\u002Fdiv>\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Cspan>\u003Ch2>📢 News\u003C\u002Fh2>\n\u003Cul>\n\u003Cli>\u003Cstrong>[2026-07-13]\u003C\u002Fstrong>: 🎉🎉🎉LightMem has been selected for the CCF ODTC open source incentive program.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>[2026-04-24]\u003C\u002Fstrong>: 🚀 LightMem now supports the latest \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fzjunlp\u002Flightmem\u002Fblob\u002FHEAD\u002Fsrc\u002Flightmem\u002Fconfigs\u002Fmemory_manager\u002Fbase_config.py\" rel=\"nofollow ugc noopener\">\u003Cstrong>DeepSeek\u003C\u002Fstrong>\u003C\u002Fa> models, including \u003Ccode>deepseek-v4-flash\u003C\u002Fcode> and \u003Ccode>deepseek-v4-pro\u003C\u002Fcode>, with \u003Ccode>reasoning_effort\u003C\u002Fcode> and thinking-mode configuration!\u003C\u002Fli>\n\u003Cli>\u003Cstrong>[2026-04-24]\u003C\u002Fstrong>: 🎉🎉🎉 \u003Ca href=\"https:\u002F\u002Farxiv.org\u002Fabs\u002F2604.21748\" rel=\"nofollow ugc noopener\">\u003Cstrong>StructMem: Structured Memory for Long-Horizon Behavior in LLMs\u003C\u002Fstrong>\u003C\u002Fa> has been accepted by \u003Cstrong>ACL 2026\u003C\u002Fstrong>!\u003C\u002Fli>\n\u003Cli>\u003Cstrong>[2026-03-21]\u003C\u002Fstrong>: 🚀 We provide a more comprehensive \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fzjunlp\u002FMemBase\" rel=\"nofollow ugc noopener\">baseline evaluation framework\u003C\u002Fa>, supporting the benchmarking of memory layers such as Mem0, A-MEM, EverMemOS, LangMem on multiple datasets like LoCoMo and LongMemEval.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>[2026-02-15]\u003C\u002Fstrong>: 🚀 \u003Cstrong>\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fzjunlp\u002Flightmem\u002Fblob\u002FHEAD\u002FStructMem.md\" rel=\"nofollow ugc noopener\">StructMem\u003C\u002Fa>\u003C\u002Fstrong> is released: A hierarchical memory framework that preserves event-level memory bindings and cross-event memory connections. \u003C\u002Fli>\n\u003Cli>\u003Cstrong>[2026-01-26]\u003C\u002Fstrong>: 🎉🎉🎉 \u003Ca href=\"https:\u002F\u002Farxiv.org\u002Fabs\u002F2510.18866\" rel=\"nofollow ugc noopener\">\u003Cstrong>LightMem: Lightweight and Efficient Memory-Augmented Generation\u003C\u002Fstrong>\u003C\u002Fa> has been accepted by \u003Cstrong>ICLR 2026\u003C\u002Fstrong>!\u003C\u002Fli>\n\u003Cli>\u003Cstrong>[2026-01-17]\u003C\u002Fstrong>: 🚀 We provide a comprehensive \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fzjunlp\u002FLightMem\u002Fblob\u002Fmain\u002Fsrc\u002Flightmem\u002Fmemory_toolkits\u002Freadme.md\" rel=\"nofollow ugc noopener\">baseline evaluation framework\u003C\u002Fa>, supporting the benchmarking of memory layers such as Mem0, A-MEM, and LangMem on multiple datasets like LoCoMo and LongMemEval.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>[2025-12-09]\u003C\u002Fstrong>: 🎬 Released a \u003Cstrong>\u003Ca href=\"#demo\" rel=\"nofollow ugc noopener\">Demo Video\u003C\u002Fa>\u003C\u002Fstrong> showcasing long-context handling, along with comprehensive \u003Cstrong>\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fzjunlp\u002Flightmem\u002Fblob\u002FHEAD\u002Ftutorial-notebooks\u002F\" rel=\"nofollow ugc noopener\">Tutorial Notebooks\u003C\u002Fa>\u003C\u002Fstrong> for various scenarios!\u003C\u002Fli>\n\u003Cli>\u003Cstrong>[2025-11-30]\u003C\u002Fstrong>: 🚌 LightMem now supports calling multiple tools provided by its \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fzjunlp\u002FLightMem\u002Fblob\u002Fmain\u002Fmcp\u002Fserver.py\" rel=\"nofollow ugc noopener\">\u003Cstrong>MCP Server\u003C\u002Fstrong>\u003C\u002Fa>.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>[2025-11-26]\u003C\u002Fstrong>: 🚀 Added full \u003Cstrong>LoCoMo\u003C\u002Fstrong> dataset support, delivering strong \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fzjunlp\u002FLightMem?tab=readme-ov-file#locomo\" rel=\"nofollow ugc noopener\">results\u003C\u002Fa> with leading performance and efficiency! Here is the [\u003Cstrong>reproduction script\u003C\u002Fstrong>](\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fzjunlp\u002FLightMem\u002Fblob\u002Fmain\u002Fexperiments\u002Flocomo\u002Fread\" rel=\"nofollow ugc noopener\">https:\u002F\u002Fgithub.com\u002Fzjunlp\u002FLightMem\u002Fblob\u002Fmain\u002Fexperiments\u002Flocomo\u002Fread\u003C\u002Fa>\u003C\u002Fli>\n\u003C\u002Ful>\n\u003C\u002Fspan>",1784564568099]