[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"similar-MemTensor--MemOS":3,"tool-MemTensor--MemOS":62},[4,18,26,36,46,54],{"id":5,"name":6,"github_repo":7,"description_zh":8,"stars":9,"difficulty_score":10,"last_commit_at":11,"category_tags":12,"status":17},4358,"openclaw","openclaw\u002Fopenclaw","OpenClaw 是一款专为个人打造的本地化 AI 助手，旨在让你在自己的设备上拥有完全可控的智能伙伴。它打破了传统 AI 助手局限于特定网页或应用的束缚，能够直接接入你日常使用的各类通讯渠道，包括微信、WhatsApp、Telegram、Discord、iMessage 等数十种平台。无论你在哪个聊天软件中发送消息，OpenClaw 都能即时响应，甚至支持在 macOS、iOS 和 Android 设备上进行语音交互，并提供实时的画布渲染功能供你操控。\n\n这款工具主要解决了用户对数据隐私、响应速度以及“始终在线”体验的需求。通过将 AI 部署在本地，用户无需依赖云端服务即可享受快速、私密的智能辅助，真正实现了“你的数据，你做主”。其独特的技术亮点在于强大的网关架构，将控制平面与核心助手分离，确保跨平台通信的流畅性与扩展性。\n\nOpenClaw 非常适合希望构建个性化工作流的技术爱好者、开发者，以及注重隐私保护且不愿被单一生态绑定的普通用户。只要具备基础的终端操作能力（支持 macOS、Linux 及 Windows WSL2），即可通过简单的命令行引导完成部署。如果你渴望拥有一个懂你",349277,3,"2026-04-06T06:32:30",[13,14,15,16],"Agent","开发框架","图像","数据工具","ready",{"id":19,"name":20,"github_repo":21,"description_zh":22,"stars":23,"difficulty_score":10,"last_commit_at":24,"category_tags":25,"status":17},3808,"stable-diffusion-webui","AUTOMATIC1111\u002Fstable-diffusion-webui","stable-diffusion-webui 是一个基于 Gradio 构建的网页版操作界面，旨在让用户能够轻松地在本地运行和使用强大的 Stable Diffusion 图像生成模型。它解决了原始模型依赖命令行、操作门槛高且功能分散的痛点，将复杂的 AI 绘图流程整合进一个直观易用的图形化平台。\n\n无论是希望快速上手的普通创作者、需要精细控制画面细节的设计师，还是想要深入探索模型潜力的开发者与研究人员，都能从中获益。其核心亮点在于极高的功能丰富度：不仅支持文生图、图生图、局部重绘（Inpainting）和外绘（Outpainting）等基础模式，还独创了注意力机制调整、提示词矩阵、负向提示词以及“高清修复”等高级功能。此外，它内置了 GFPGAN 和 CodeFormer 等人脸修复工具，支持多种神经网络放大算法，并允许用户通过插件系统无限扩展能力。即使是显存有限的设备，stable-diffusion-webui 也提供了相应的优化选项，让高质量的 AI 艺术创作变得触手可及。",162132,"2026-04-05T11:01:52",[14,15,13],{"id":27,"name":28,"github_repo":29,"description_zh":30,"stars":31,"difficulty_score":32,"last_commit_at":33,"category_tags":34,"status":17},1381,"everything-claude-code","affaan-m\u002Feverything-claude-code","everything-claude-code 是一套专为 AI 编程助手（如 Claude Code、Codex、Cursor 等）打造的高性能优化系统。它不仅仅是一组配置文件，而是一个经过长期实战打磨的完整框架，旨在解决 AI 代理在实际开发中面临的效率低下、记忆丢失、安全隐患及缺乏持续学习能力等核心痛点。\n\n通过引入技能模块化、直觉增强、记忆持久化机制以及内置的安全扫描功能，everything-claude-code 能显著提升 AI 在复杂任务中的表现，帮助开发者构建更稳定、更智能的生产级 AI 代理。其独特的“研究优先”开发理念和针对 Token 消耗的优化策略，使得模型响应更快、成本更低，同时有效防御潜在的攻击向量。\n\n这套工具特别适合软件开发者、AI 研究人员以及希望深度定制 AI 工作流的技术团队使用。无论您是在构建大型代码库，还是需要 AI 协助进行安全审计与自动化测试，everything-claude-code 都能提供强大的底层支持。作为一个曾荣获 Anthropic 黑客大奖的开源项目，它融合了多语言支持与丰富的实战钩子（hooks），让 AI 真正成长为懂上",158594,2,"2026-04-16T23:34:05",[14,13,35],"语言模型",{"id":37,"name":38,"github_repo":39,"description_zh":40,"stars":41,"difficulty_score":42,"last_commit_at":43,"category_tags":44,"status":17},8272,"opencode","anomalyco\u002Fopencode","OpenCode 是一款开源的 AI 编程助手（Coding Agent），旨在像一位智能搭档一样融入您的开发流程。它不仅仅是一个代码补全插件，而是一个能够理解项目上下文、自主规划任务并执行复杂编码操作的智能体。无论是生成全新功能、重构现有代码，还是排查难以定位的 Bug，OpenCode 都能通过自然语言交互高效完成，显著减少开发者在重复性劳动和上下文切换上的时间消耗。\n\n这款工具专为软件开发者、工程师及技术研究人员设计，特别适合希望利用大模型能力来提升编码效率、加速原型开发或处理遗留代码维护的专业人群。其核心亮点在于完全开源的架构，这意味着用户可以审查代码逻辑、自定义行为策略，甚至私有化部署以保障数据安全，彻底打破了传统闭源 AI 助手的“黑盒”限制。\n\n在技术体验上，OpenCode 提供了灵活的终端界面（Terminal UI）和正在测试中的桌面应用程序，支持 macOS、Windows 及 Linux 全平台。它兼容多种包管理工具，安装便捷，并能无缝集成到现有的开发环境中。无论您是追求极致控制权的资深极客，还是渴望提升产出的独立开发者，OpenCode 都提供了一个透明、可信",144296,1,"2026-04-16T14:50:03",[13,45],"插件",{"id":47,"name":48,"github_repo":49,"description_zh":50,"stars":51,"difficulty_score":32,"last_commit_at":52,"category_tags":53,"status":17},2271,"ComfyUI","Comfy-Org\u002FComfyUI","ComfyUI 是一款功能强大且高度模块化的视觉 AI 引擎，专为设计和执行复杂的 Stable Diffusion 图像生成流程而打造。它摒弃了传统的代码编写模式，采用直观的节点式流程图界面，让用户通过连接不同的功能模块即可构建个性化的生成管线。\n\n这一设计巧妙解决了高级 AI 绘图工作流配置复杂、灵活性不足的痛点。用户无需具备编程背景，也能自由组合模型、调整参数并实时预览效果，轻松实现从基础文生图到多步骤高清修复等各类复杂任务。ComfyUI 拥有极佳的兼容性，不仅支持 Windows、macOS 和 Linux 全平台，还广泛适配 NVIDIA、AMD、Intel 及苹果 Silicon 等多种硬件架构，并率先支持 SDXL、Flux、SD3 等前沿模型。\n\n无论是希望深入探索算法潜力的研究人员和开发者，还是追求极致创作自由度的设计师与资深 AI 绘画爱好者，ComfyUI 都能提供强大的支持。其独特的模块化架构允许社区不断扩展新功能，使其成为当前最灵活、生态最丰富的开源扩散模型工具之一，帮助用户将创意高效转化为现实。",108322,"2026-04-10T11:39:34",[14,15,13],{"id":55,"name":56,"github_repo":57,"description_zh":58,"stars":59,"difficulty_score":32,"last_commit_at":60,"category_tags":61,"status":17},6121,"gemini-cli","google-gemini\u002Fgemini-cli","gemini-cli 是一款由谷歌推出的开源 AI 命令行工具，它将强大的 Gemini 大模型能力直接集成到用户的终端环境中。对于习惯在命令行工作的开发者而言，它提供了一条从输入提示词到获取模型响应的最短路径，无需切换窗口即可享受智能辅助。\n\n这款工具主要解决了开发过程中频繁上下文切换的痛点，让用户能在熟悉的终端界面内直接完成代码理解、生成、调试以及自动化运维任务。无论是查询大型代码库、根据草图生成应用，还是执行复杂的 Git 操作，gemini-cli 都能通过自然语言指令高效处理。\n\n它特别适合广大软件工程师、DevOps 人员及技术研究人员使用。其核心亮点包括支持高达 100 万 token 的超长上下文窗口，具备出色的逻辑推理能力；内置 Google 搜索、文件操作及 Shell 命令执行等实用工具；更独特的是，它支持 MCP（模型上下文协议），允许用户灵活扩展自定义集成，连接如图像生成等外部能力。此外，个人谷歌账号即可享受免费的额度支持，且项目基于 Apache 2.0 协议完全开源，是提升终端工作效率的理想助手。",100752,"2026-04-10T01:20:03",[45,13,15,14],{"id":63,"github_repo":64,"name":65,"description_en":66,"description_zh":67,"ai_summary_zh":67,"readme_en":68,"readme_zh":69,"quickstart_zh":70,"use_case_zh":71,"hero_image_url":72,"owner_login":73,"owner_name":74,"owner_avatar_url":75,"owner_bio":76,"owner_company":77,"owner_location":77,"owner_email":77,"owner_twitter":77,"owner_website":77,"owner_url":78,"languages":79,"stars":118,"forks":119,"last_commit_at":120,"license":121,"difficulty_score":10,"env_os":122,"env_gpu":123,"env_ram":124,"env_deps":125,"category_tags":134,"github_topics":136,"view_count":32,"oss_zip_url":77,"oss_zip_packed_at":77,"status":17,"created_at":154,"updated_at":155,"faqs":156,"releases":186},8298,"MemTensor\u002FMemOS","MemOS","AI memory OS for LLM and Agent systems(moltbot,clawdbot,openclaw), enabling persistent Skill memory for cross-task skill reuse and evolution.","MemOS 是一款专为大语言模型（LLM）和智能体系统打造的“记忆操作系统”。它旨在解决当前 AI 在跨任务场景中难以保留长期记忆、无法复用既有技能以及上下文窗口成本过高的问题。通过引入持久化的技能记忆机制，MemOS 让 AI 能够像人类一样积累经验，实现技能的持续进化与跨任务调用。\n\n该项目特别适合 AI 开发者、研究人员以及希望构建具备长期记忆能力智能体的技术团队使用。无论是需要多智能体协作的复杂应用，还是追求数据隐私的本地部署场景，MemOS 都能提供灵活支持。其核心技术亮点包括：混合搜索架构（结合全文检索与向量搜索）、自动任务总结与技能自升级机制，以及云边端协同的记忆管理方案。实测数据显示，相比传统方案，MemOS 不仅能将记忆相关的 Token 消耗降低约 35%，还在多项长程记忆基准测试中显著提升了准确率。此外，它支持完全本地化部署，确保数据不出域，同时也提供云端服务以实现多实例间的记忆共享与无缝上下文切换，是构建下一代个性化、高智商 AI 助手的关键基础设施。","\u003Cdiv align=\"center\">\n  \u003Ca href=\"https:\u002F\u002Fmemos.openmem.net\u002F\">\n    \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FMemTensor_MemOS_readme_c00b10a88cc8.gif\" alt=\"MemOS Banner\">\n  \u003C\u002Fa>\n\n  \u003Ch1 align=\"center\">\n    \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FMemTensor_MemOS_readme_5fcdd8e6a8bd.png\" alt=\"MemOS Logo\" width=\"50\"\u002F>\n    MemOS 2.0: 星尘（Stardust）\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fstatus-Preview-blue\" alt=\"Preview Badge\"\u002F>\n  \u003C\u002Fh1>\n\n  \u003Cp>\n    \u003Ca href=\"https:\u002F\u002Fwww.memtensor.com.cn\u002F\">\n      \u003Cimg alt=\"Static Badge\" src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FMaintained_by-MemTensor-blue\">\n    \u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fpypi.org\u002Fproject\u002FMemoryOS\">\n      \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fpypi\u002Fv\u002FMemoryOS?label=pypi%20package\" alt=\"PyPI Version\">\n    \u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fpypi.org\u002Fproject\u002FMemoryOS\">\n      \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fpypi\u002Fpyversions\u002FMemoryOS.svg\" alt=\"Supported Python versions\">\n    \u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fpypi.org\u002Fproject\u002FMemoryOS\">\n      \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FPlatform-Linux%20%7C%20macOS%20%7C%20Windows-lightgrey\" alt=\"Supported Platforms\">\n    \u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fmemos-docs.openmem.net\u002Fhome\u002Foverview\u002F\">\n      \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FDocumentation-view-blue.svg\" alt=\"Documentation\">\n    \u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Farxiv.org\u002Fabs\u002F2507.03724\">\n      \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FarXiv-2507.03724-b31b1b.svg\" alt=\"ArXiv Paper\">\n    \u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fdiscussions\">\n      \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FGitHub-Discussions-181717.svg?logo=github\" alt=\"GitHub Discussions\">\n    \u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fdiscord.gg\u002FTxbx3gebZR\">\n      \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FDiscord-join%20chat-7289DA.svg?logo=discord\" alt=\"Discord\">\n    \u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FMemTensor_MemOS_readme_423b366f315e.png\">\n      \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FWeChat-Group-07C160.svg?logo=wechat\" alt=\"WeChat Group\">\n    \u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fopensource.org\u002Flicense\u002Fapache-2-0\u002F\">\n      \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FLicense-Apache_2.0-green.svg?logo=apache\" alt=\"License\">\n    \u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fgithub.com\u002FIAAR-Shanghai\u002FAwesome-AI-Memory\">\n      \u003Cimg alt=\"Awesome AI Memory\" src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FResources-Awesome--AI--Memory-8A2BE2\">\n    \u003C\u002Fa>\n  \u003C\u002Fp>\n\n\u003Cp align=\"center\">\n  \u003Cstrong>🎯 +43.70% Accuracy vs. OpenAI Memory\u003C\u002Fstrong>\u003Cbr\u002F>\n  \u003Cstrong>🏆 Top-tier long-term memory + personalization\u003C\u002Fstrong>\u003Cbr\u002F>\n  \u003Cstrong>💰 Saves 35.24% memory tokens\u003C\u002Fstrong>\u003Cbr\u002F>\n  \u003Csub>LoCoMo 75.80 • LongMemEval +40.43% • PrefEval-10 +2568% • PersonaMem +40.75%\u003C\u002Fsub>\n  \u003C!-- \u003Ca href=\"https:\u002F\u002Fmemos.openmem.net\u002F\">\n    \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FMemTensor_MemOS_readme_9f72e921e59d.gif\" alt=\"MemOS Free API Banner\">\n  \u003C\u002Fa> -->\n\n\u003C\u002Fp>\n\n\u003C\u002Fdiv>\n\n\u003C!-- Get Free API: [Try API](https:\u002F\u002Fmemos-dashboard.openmem.net\u002Fquickstart\u002F?source=github) -->\n\n\u003C!-- --- -->\n\n\u003C!-- \u003Cbr> -->\n\n## 🦞 Enhanced OpenClaw with MemOS Plugin\n\n![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FMemTensor_MemOS_readme_2c4094df8572.png)\n\n🦞 Your lobster now has a working memory system — choose **Cloud** or **Local** to get started.\n\n### ☁️ Cloud Plugin — Hosted Memory Service\n\n- [**72% lower token usage**](https:\u002F\u002Fx.com\u002FMemOS_dev\u002Fstatus\u002F2020854044583924111) — intelligent memory retrieval instead of loading full chat history\n- [**Multi-agent memory sharing**](https:\u002F\u002Fx.com\u002FMemOS_dev\u002Fstatus\u002F2020538135487062094) — multi-instance agents share memory via same user_id, automatic context handoff\n\nGet your API key: [MemOS Dashboard](https:\u002F\u002Fmemos-dashboard.openmem.net\u002Fcn\u002Flogin\u002F)\nFull tutorial → [MemOS-Cloud-OpenClaw-Plugin](https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS-Cloud-OpenClaw-Plugin)\n\n### 🧠 Local Plugin — 100% On-Device Memory\n\n- **Zero cloud dependency** — all data stays on your machine, persistent local SQLite storage\n- **Hybrid search + task & skill evolution** — FTS5 + vector search, auto task summarization, reusable skills that self-upgrade\n- **Multi-agent collaboration + Memory Viewer** — memory isolation, skill sharing, full web dashboard with 7 management pages\n\n 🌐 [Homepage](https:\u002F\u002Fmemos-claw.openmem.net) ·\n📖 [Documentation](https:\u002F\u002Fmemos-claw.openmem.net\u002Fdocs\u002Findex.html) · 📦 [NPM](https:\u002F\u002Fwww.npmjs.com\u002Fpackage\u002F@memtensor\u002Fmemos-local-openclaw-plugin)\n\n## 📌 MemOS: Memory Operating System for AI Agents\n\n**MemOS** is a Memory Operating System for LLMs and AI agents that unifies **store \u002F retrieve \u002F manage** for long-term memory, enabling **context-aware and personalized** interactions with **KB**, **multi-modal**, **tool memory**, and **enterprise-grade** optimizations built in.\n\n\n\n### Key Features\n\n- **Unified Memory API**: A single API to add, retrieve, edit, and delete memory—structured as a graph, inspectable and editable by design, not a black-box embedding store.\n- **Multi-Modal Memory**: Natively supports text, images, tool traces, and personas, retrieved and reasoned together in one memory system.\n- **Multi-Cube Knowledge Base Management**: Manage multiple knowledge bases as composable memory cubes, enabling isolation, controlled sharing, and dynamic composition across users, projects, and agents.\n- **Asynchronous Ingestion via MemScheduler**: Run memory operations asynchronously with millisecond-level latency for production stability under high concurrency.\n- **Memory Feedback & Correction**: Refine memory with natural-language feedback—correcting, supplementing, or replacing existing memories over time.\n\n\n### News\n\n- **2026-03-08** · 🦞 **MemOS OpenClaw Plugin — Cloud & Local**\n  Official OpenClaw memory plugins launched. **Cloud Plugin**: hosted memory service with 72% lower token usage and multi-agent memory sharing ([MemOS-Cloud-OpenClaw-Plugin](https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS-Cloud-OpenClaw-Plugin)). **Local Plugin** (`v1.0.0`): 100% on-device memory with persistent SQLite, hybrid search (FTS5 + vector), task summarization & skill evolution, multi-agent collaboration, and a full Memory Viewer dashboard.\n\n- **2025-12-24** · 🎉 **MemOS v2.0: Stardust (星尘) Release**\n  Comprehensive KB (doc\u002FURL parsing + cross-project sharing), memory feedback & precise deletion, multi-modal memory (images\u002Fcharts), tool memory for agent planning, Redis Streams scheduling + DB optimizations, streaming\u002Fnon-streaming chat, MCP upgrade, and lightweight quick\u002Ffull deployment.\n  \u003Cdetails>\n    \u003Csummary>✨ \u003Cb>New Features\u003C\u002Fb>\u003C\u002Fsummary>\n\n  **Knowledge Base & Memory**\n  - Added knowledge base support for long-term memory from documents and URLs\n\n  **Feedback & Memory Management**\n  - Added natural language feedback and correction for memories\n  - Added memory deletion API by memory ID\n  - Added MCP support for memory deletion and feedback\n\n  **Conversation & Retrieval**\n  - Added chat API with memory-aware retrieval\n  - Added memory filtering with custom tags (Cloud & Open Source)\n\n  **Multimodal & Tool Memory**\n  - Added tool memory for tool usage history\n  - Added image memory support for conversations and documents\n\n  \u003C\u002Fdetails>\n\n  \u003Cdetails>\n    \u003Csummary>📈 \u003Cb>Improvements\u003C\u002Fb>\u003C\u002Fsummary>\n\n  **Data & Infrastructure**\n  - Upgraded database for better stability and performance\n\n  **Scheduler**\n  - Rebuilt task scheduler with Redis Streams and queue isolation\n  - Added task priority, auto-recovery, and quota-based scheduling\n\n  **Deployment & Engineering**\n  - Added lightweight deployment with quick and full modes\n\n  \u003C\u002Fdetails>\n\n  \u003Cdetails>\n    \u003Csummary>🐞 \u003Cb>Bug Fixes\u003C\u002Fb>\u003C\u002Fsummary>\n\n  **Memory Scheduling & Updates**\n  - Fixed legacy scheduling API to ensure correct memory isolation\n  - Fixed memory update logging to show new memories correctly\n\n  \u003C\u002Fdetails>\n\n- **2025-08-07** · 🎉 **MemOS v1.0.0 (MemCube) Release**\n  First MemCube release with a word-game demo, LongMemEval evaluation, BochaAISearchRetriever integration, improved search capabilities, and the official Playground launch.\n\n  \u003Cdetails>\n    \u003Csummary>✨ \u003Cb>New Features\u003C\u002Fb>\u003C\u002Fsummary>\n\n  **Playground**\n  - Expanded Playground features and algorithm performance.\n\n  **MemCube Construction**\n  - Added a text game demo based on the MemCube novel.\n\n  **Extended Evaluation Set**\n  - Added LongMemEval evaluation results and scripts.\n\n  \u003C\u002Fdetails>\n\n  \u003Cdetails>\n    \u003Csummary>📈 \u003Cb>Improvements\u003C\u002Fb>\u003C\u002Fsummary>\n\n  **Plaintext Memory**\n  - Integrated internet search with Bocha.\n  - Expanded graph database support.\n  - Added contextual understanding for the tree-structured plaintext memory search interface.\n\n  \u003C\u002Fdetails>\n\n  \u003Cdetails>\n    \u003Csummary>🐞 \u003Cb>Bug Fixes\u003C\u002Fb>\u003C\u002Fsummary>\n\n  **KV Cache Concatenation**\n  - Fixed the concat_cache method.\n\n  **Plaintext Memory**\n  - Fixed graph search-related issues.\n\n  \u003C\u002Fdetails>\n\n- **2025-07-07** · 🎉 **MemOS v1.0: Stellar (星河) Preview Release**\n  A SOTA Memory OS for LLMs is now open-sourced.\n- **2025-07-04** · 🎉 **MemOS Paper Release**\n  [MemOS: A Memory OS for AI System](https:\u002F\u002Farxiv.org\u002Fabs\u002F2507.03724) is available on arXiv.\n- **2024-07-04** · 🎉 **Memory3 Model Release at WAIC 2024**\n  The Memory3 model, featuring a memory-layered architecture, was unveiled at the 2024 World Artificial Intelligence Conference.\n\n\u003Cbr>\n\n## 🚀 Quickstart Guide\n\n### ☁️ 1、Cloud API (Hosted)\n#### Get API Key\n- Sign up on the [MemOS dashboard](https:\u002F\u002Fmemos-dashboard.openmem.net\u002Fcn\u002Fquickstart\u002F?source=landing)\n- Go to **API Keys** and copy your key\n\n#### Next Steps\n- [MemOS Cloud Getting Started](https:\u002F\u002Fmemos-docs.openmem.net\u002Fmemos_cloud\u002Fquick_start\u002F)\n  Connect to MemOS Cloud and enable memory in minutes.\n- [MemOS Cloud Platform](https:\u002F\u002Fmemos.openmem.net\u002F?from=\u002Fquickstart\u002F)\n  Explore the Cloud dashboard, features, and workflows.\n\n### 🖥️ 2、Self-Hosted (Local\u002FPrivate)\n1. Get the repository.\n    ```bash\n    git clone https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS.git\n    cd MemOS\n    pip install -r .\u002Fdocker\u002Frequirements.txt\n    ```\n2. Configure `docker\u002F.env.example` and copy to `MemOS\u002F.env`\n - The `OPENAI_API_KEY`,`MOS_EMBEDDER_API_KEY`,`MEMRADER_API_KEY` and others can be applied for through [`BaiLian`](https:\u002F\u002Fbailian.console.aliyun.com\u002F?spm=a2c4g.11186623.0.0.2f2165b08fRk4l&tab=api#\u002Fapi).\n - Fill in the corresponding configuration in the `MemOS\u002F.env` file.\n - Supported LLM providers: **OpenAI**, **Azure OpenAI**, **Qwen (DashScope)**, **DeepSeek**, **MiniMax**, **Ollama**, **HuggingFace**, **vLLM**. Set `MOS_CHAT_MODEL_PROVIDER` to select the backend (e.g., `openai`, `qwen`, `deepseek`, `minimax`).\n3. Start the service.\n\n- Launch via Docker\n  ###### Tips: Please ensure that Docker Compose is installed successfully and that you have navigated to the docker directory (via `cd docker`) before executing the following command.\n  ```bash\n  # Enter docker directory\n  docker compose up\n  ```\n  ##### For detailed steps, see the[`Docker Reference`](https:\u002F\u002Fdocs.openmem.net\u002Fopen_source\u002Fgetting_started\u002Frest_api_server\u002F#method-1-docker-use-repository-dependency-package-imagestart-recommended-use).\n\n- Launch via the uvicorn command line interface (CLI)\n  ###### Tips: Please ensure that Neo4j and Qdrant are running before executing the following command.\n  ```bash\n  cd src\n  uvicorn memos.api.server_api:app --host 0.0.0.0 --port 8001 --workers 1\n  ```\n  ##### For detailed integration steps, see the [`CLI Reference`](https:\u002F\u002Fdocs.openmem.net\u002Fopen_source\u002Fgetting_started\u002Frest_api_server\u002F#method-3client-install-with-CLI).\n\n\n\n### Basic Usage (Self-Hosted)\n  - Add User Message\n    ```python\n    import requests\n    import json\n\n    data = {\n        \"user_id\": \"8736b16e-1d20-4163-980b-a5063c3facdc\",\n        \"mem_cube_id\": \"b32d0977-435d-4828-a86f-4f47f8b55bca\",\n        \"messages\": [\n            {\n                \"role\": \"user\",\n                \"content\": \"I like strawberry\"\n            }\n        ],\n        \"async_mode\": \"sync\"\n    }\n    headers = {\n        \"Content-Type\": \"application\u002Fjson\"\n    }\n    url = \"http:\u002F\u002Flocalhost:8000\u002Fproduct\u002Fadd\"\n\n    res = requests.post(url=url, headers=headers, data=json.dumps(data))\n    print(f\"result: {res.json()}\")\n    ```\n  - Search User Memory\n    ```python\n    import requests\n    import json\n\n    data = {\n        \"query\": \"What do I like\",\n        \"user_id\": \"8736b16e-1d20-4163-980b-a5063c3facdc\",\n        \"mem_cube_id\": \"b32d0977-435d-4828-a86f-4f47f8b55bca\"\n    }\n    headers = {\n        \"Content-Type\": \"application\u002Fjson\"\n    }\n    url = \"http:\u002F\u002Flocalhost:8000\u002Fproduct\u002Fsearch\"\n\n    res = requests.post(url=url, headers=headers, data=json.dumps(data))\n    print(f\"result: {res.json()}\")\n    ```\n\n\u003Cbr>\n\n## 📚 Resources\n\n- **Awesome-AI-Memory**\n This is a curated repository dedicated to resources on memory and memory systems for large language models. It systematically collects relevant research papers, frameworks, tools, and practical insights. The repository aims to organize and present the rapidly evolving research landscape of LLM memory, bridging multiple research directions including natural language processing, information retrieval, agentic systems, and cognitive science.\n- **Get started** 👉 [IAAR-Shanghai\u002FAwesome-AI-Memory](https:\u002F\u002Fgithub.com\u002FIAAR-Shanghai\u002FAwesome-AI-Memory)\n- **MemOS Cloud OpenClaw Plugin**\n  Official OpenClaw lifecycle plugin for MemOS Cloud. It automatically recalls context from MemOS before the agent starts and saves the conversation back to MemOS after the agent finishes.\n- **Get started** 👉 [MemTensor\u002FMemOS-Cloud-OpenClaw-Plugin](https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS-Cloud-OpenClaw-Plugin)\n\n\u003Cbr>\n\n## 💬 Community & Support\n\nJoin our community to ask questions, share your projects, and connect with other developers.\n\n- **GitHub Issues**: Report bugs or request features in our \u003Ca href=\"https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fissues\" target=\"_blank\">GitHub Issues\u003C\u002Fa>.\n- **GitHub Pull Requests**: Contribute code improvements via \u003Ca href=\"https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpulls\" target=\"_blank\">Pull Requests\u003C\u002Fa>.\n- **GitHub Discussions**: Participate in our \u003Ca href=\"https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fdiscussions\" target=\"_blank\">GitHub Discussions\u003C\u002Fa> to ask questions or share ideas.\n- **Discord**: Join our \u003Ca href=\"https:\u002F\u002Fdiscord.gg\u002FTxbx3gebZR\" target=\"_blank\">Discord Server\u003C\u002Fa>.\n- **WeChat**: Scan the QR code to join our WeChat group.\n\n\u003Cdiv align=\"center\">\n  \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FMemTensor_MemOS_readme_423b366f315e.png\" alt=\"QR Code\" width=\"300\" \u002F>\n\u003C\u002Fdiv>\n\n\u003Cbr>\n\n## 📜 Citation\n\n> [!NOTE]\n> We publicly released the Short Version on **May 28, 2025**, making it the earliest work to propose the concept of a Memory Operating System for LLMs.\n\nIf you use MemOS in your research, we would appreciate citations to our papers.\n\n```bibtex\n\n@article{li2025memos_long,\n  title={MemOS: A Memory OS for AI System},\n  author={Li, Zhiyu and Song, Shichao and Xi, Chenyang and Wang, Hanyu and Tang, Chen and Niu, Simin and Chen, Ding and Yang, Jiawei and Li, Chunyu and Yu, Qingchen and Zhao, Jihao and Wang, Yezhaohui and Liu, Peng and Lin, Zehao and Wang, Pengyuan and Huo, Jiahao and Chen, Tianyi and Chen, Kai and Li, Kehang and Tao, Zhen and Ren, Junpeng and Lai, Huayi and Wu, Hao and Tang, Bo and Wang, Zhenren and Fan, Zhaoxin and Zhang, Ningyu and Zhang, Linfeng and Yan, Junchi and Yang, Mingchuan and Xu, Tong and Xu, Wei and Chen, Huajun and Wang, Haofeng and Yang, Hongkang and Zhang, Wentao and Xu, Zhi-Qin John and Chen, Siheng and Xiong, Feiyu},\n  journal={arXiv preprint arXiv:2507.03724},\n  year={2025},\n  url={https:\u002F\u002Farxiv.org\u002Fabs\u002F2507.03724}\n}\n\n@article{li2025memos_short,\n  title={MemOS: An Operating System for Memory-Augmented Generation (MAG) in Large Language Models},\n  author={Li, Zhiyu and Song, Shichao and Wang, Hanyu and Niu, Simin and Chen, Ding and Yang, Jiawei and Xi, Chenyang and Lai, Huayi and Zhao, Jihao and Wang, Yezhaohui and others},\n  journal={arXiv preprint arXiv:2505.22101},\n  year={2025},\n  url={https:\u002F\u002Farxiv.org\u002Fabs\u002F2505.22101}\n}\n\n@article{yang2024memory3,\nauthor = {Yang, Hongkang and Zehao, Lin and Wenjin, Wang and Wu, Hao and Zhiyu, Li and Tang, Bo and Wenqiang, Wei and Wang, Jinbo and Zeyun, Tang and Song, Shichao and Xi, Chenyang and Yu, Yu and Kai, Chen and Xiong, Feiyu and Tang, Linpeng and Weinan, E},\ntitle = {Memory$^3$: Language Modeling with Explicit Memory},\njournal = {Journal of Machine Learning},\nyear = {2024},\nvolume = {3},\nnumber = {3},\npages = {300--346},\nissn = {2790-2048},\ndoi = {https:\u002F\u002Fdoi.org\u002F10.4208\u002Fjml.240708},\nurl = {https:\u002F\u002Fglobal-sci.com\u002Farticle\u002F91443\u002Fmemory3-language-modeling-with-explicit-memory}\n}\n```\n\n\u003Cbr>\n\n## 🙌 Contributing\n\nWe welcome contributions from the community! Please read our [contribution guidelines](https:\u002F\u002Fmemos-docs.openmem.net\u002Fopen_source\u002Fcontribution\u002Foverview\u002F) to get started.\n\n\u003Cbr>\n\n## 📄 License\n\nMemOS is licensed under the [Apache 2.0 License](.\u002FLICENSE).\n","\u003Cdiv align=\"center\">\n  \u003Ca href=\"https:\u002F\u002Fmemos.openmem.net\u002F\">\n    \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FMemTensor_MemOS_readme_c00b10a88cc8.gif\" alt=\"MemOS Banner\">\n  \u003C\u002Fa>\n\n  \u003Ch1 align=\"center\">\n    \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FMemTensor_MemOS_readme_5fcdd8e6a8bd.png\" alt=\"MemOS Logo\" width=\"50\"\u002F>\n    MemOS 2.0：星尘（Stardust）\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fstatus-Preview-blue\" alt=\"Preview Badge\"\u002F>\n  \u003C\u002Fh1>\n\n  \u003Cp>\n    \u003Ca href=\"https:\u002F\u002Fwww.memtensor.com.cn\u002F\">\n      \u003Cimg alt=\"Static Badge\" src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FMaintained_by-MemTensor-blue\">\n    \u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fpypi.org\u002Fproject\u002FMemoryOS\">\n      \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fpypi\u002Fv\u002FMemoryOS?label=pypi%20package\" alt=\"PyPI Version\">\n    \u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fpypi.org\u002Fproject\u002FMemoryOS\">\n      \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fpypi\u002Fpyversions\u002FMemoryOS.svg\" alt=\"Supported Python versions\">\n    \u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fpypi.org\u002Fproject\u002FMemoryOS\">\n      \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FPlatform-Linux%20%7C%20macOS%20%7C%20Windows-lightgrey\" alt=\"Supported Platforms\">\n    \u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fmemos-docs.openmem.net\u002Fhome\u002Foverview\u002F\">\n      \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FDocumentation-view-blue.svg\" alt=\"Documentation\">\n    \u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Farxiv.org\u002Fabs\u002F2507.03724\">\n      \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FarXiv-2507.03724-b31b1b.svg\" alt=\"ArXiv Paper\">\n    \u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fdiscussions\">\n      \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FGitHub-Discussions-181717.svg?logo=github\" alt=\"GitHub Discussions\">\n    \u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fdiscord.gg\u002FTxbx3gebZR\">\n      \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FDiscord-join%20chat-7289DA.svg?logo=discord\" alt=\"Discord\">\n    \u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FMemTensor_MemOS_readme_423b366f315e.png\">\n      \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FWeChat-Group-07C160.svg?logo=wechat\" alt=\"WeChat Group\">\n    \u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fopensource.org\u002Flicense\u002Fapache-2-0\u002F\">\n      \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FLicense-Apache_2.0-green.svg?logo=apache\" alt=\"License\">\n    \u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fgithub.com\u002FIAAR-Shanghai\u002FAwesome-AI-Memory\">\n      \u003Cimg alt=\"Awesome AI Memory\" src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FResources-Awesome--AI--Memory-8A2BE2\">\n    \u003C\u002Fa>\n  \u003C\u002Fp>\n\n\u003Cp align=\"center\">\n  \u003Cstrong>🎯 相比 OpenAI 内存，准确率提升 43.70%\u003C\u002Fstrong>\u003Cbr\u002F>\n  \u003Cstrong>🏆 顶级的长期记忆与个性化功能\u003C\u002Fstrong>\u003Cbr\u002F>\n  \u003Cstrong>💰 节省 35.24% 的内存 token 消耗\u003C\u002Fstrong>\u003Cbr\u002F>\n  \u003Csub>LoCoMo 75.80 • LongMemEval 提升 40.43% • PrefEval-10 提升 2568% • PersonaMem 提升 40.75%\u003C\u002Fsub>\n  \u003C!-- \u003Ca href=\"https:\u002F\u002Fmemos.openmem.net\u002F\">\n    \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FMemTensor_MemOS_readme_9f72e921e59d.gif\" alt=\"MemOS 免费 API Banner\">\n  \u003C\u002Fa> -->\n\n\u003C\u002Fp>\n\n\u003C\u002Fdiv>\n\n\u003C!-- 获取免费 API：[试用 API](https:\u002F\u002Fmemos-dashboard.openmem.net\u002Fquickstart\u002F?source=github) -->\n\n\u003C!-- --- -->\n\n\u003C!-- \u003Cbr> -->\n\n## 🦞 使用 MemOS 插件增强 OpenClaw\n\n![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FMemTensor_MemOS_readme_2c4094df8572.png)\n\n🦞 您的龙虾现在拥有了工作记忆系统——选择“云端”或“本地”即可开始使用。\n\n### ☁️ 云端插件——托管式内存服务\n\n- [**token 使用量降低 72%**](https:\u002F\u002Fx.com\u002FMemOS_dev\u002Fstatus\u002F2020854044583924111)——智能检索内存，而非加载完整聊天记录\n- [**多智能体共享内存**](https:\u002F\u002Fx.com\u002FMemOS_dev\u002Fstatus\u002F2020538135487062094)——多个实例的智能体通过相同的 user_id 共享内存，自动进行上下文交接\n\n获取您的 API 密钥：[MemOS 控制台](https:\u002F\u002Fmemos-dashboard.openmem.net\u002Fcn\u002Flogin\u002F)\n完整教程 → [MemOS-Cloud-OpenClaw-Plugin](https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS-Cloud-OpenClaw-Plugin)\n\n### 🧠 本地插件——100% 本地设备内存\n\n- **零云端依赖**——所有数据均保存在您的设备上，采用持久化的本地 SQLite 存储\n- **混合搜索 + 任务与技能进化**——FTS5 + 向量搜索，自动总结任务，可重复使用的技能会自我升级\n- **多智能体协作 + 内存查看器**——内存隔离、技能共享，配备包含 7 个管理页面的完整 Web 控制台\n\n🌐 [首页](https:\u002F\u002Fmemos-claw.openmem.net) ·\n📖 [文档](https:\u002F\u002Fmemos-claw.openmem.net\u002Fdocs\u002Findex.html) · 📦 [NPM](https:\u002F\u002Fwww.npmjs.com\u002Fpackage\u002F@memtensor\u002Fmemos-local-openclaw-plugin)\n\n## 📌 MemOS：面向 AI 智能体的记忆操作系统\n\n**MemOS** 是一款专为 LLM 和 AI 智能体设计的记忆操作系统，它统一了长期记忆的**存储 \u002F 检索 \u002F 管理**功能，从而实现**情境感知与个性化**的交互，并内置了对**知识库**、**多模态信息**、**工具记忆**以及**企业级优化**的支持。\n\n\n\n### 核心特性\n\n- **统一的记忆 API**：一个单一的 API 即可完成记忆的添加、检索、编辑和删除——以图结构组织，设计上可检查和编辑，而非黑盒式的嵌入存储。\n- **多模态记忆**：原生支持文本、图像、工具调用痕迹和角色人格，可在同一记忆系统中共同检索并进行推理。\n- **多立方体知识库管理**：将多个知识库作为可组合的记忆立方体进行管理，实现隔离、受控共享以及跨用户、项目和智能体的动态组合。\n- **通过 MemScheduler 进行异步摄入**：以毫秒级延迟异步执行记忆操作，确保高并发下的生产稳定性。\n- **记忆反馈与修正**：通过自然语言反馈来优化记忆——随着时间推移，可以纠正、补充或替换现有记忆。\n\n### 新闻\n\n- **2026年3月8日** · 🦞 **MemOS OpenClaw 插件 — 云端与本地**\n  官方 OpenClaw 内存插件正式发布。**云端插件**：托管式内存服务，令牌使用量降低 72%，支持多智能体内存共享（[MemOS-Cloud-OpenClaw-Plugin](https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS-Cloud-OpenClaw-Plugin)）。**本地插件**（`v1.0.0`）：100% 本地设备端内存，采用持久化 SQLite 数据库、混合搜索（FTS5 + 向量）、任务摘要与技能进化、多智能体协作，以及完整的内存查看器仪表盘。\n\n- **2025年12月24日** · 🎉 **MemOS v2.0：Stardust（星尘）发布**\n  全面的知识库功能（文档\u002FURL 解析 + 跨项目共享）、记忆反馈与精准删除、多模态记忆（图片\u002F图表）、用于智能体规划的工具记忆、Redis Streams 调度 + 数据库优化、流式\u002F非流式聊天、MCP 升级，以及轻量级快速\u002F完整部署。\n  \u003Cdetails>\n    \u003Csummary>✨ \u003Cb>新特性\u003C\u002Fb>\u003C\u002Fsummary>\n\n  **知识库与记忆**\n  - 新增知识库支持，可从文档和 URL 中获取长期记忆\n\n  **反馈与记忆管理**\n  - 新增自然语言反馈与记忆修正功能\n  - 新增基于记忆 ID 的记忆删除 API\n  - 新增对 MCP 的支持，用于记忆删除和反馈\n\n  **对话与检索**\n  - 新增具备记忆感知检索功能的聊天 API\n  - 新增自定义标签的记忆过滤功能（云端与开源版）\n\n  **多模态与工具记忆**\n  - 新增工具记忆，用于记录工具使用历史\n  - 新增图像记忆支持，可用于对话和文档\n\n  \u003C\u002Fdetails>\n\n  \u003Cdetails>\n    \u003Csummary>📈 \u003Cb>改进\u003C\u002Fb>\u003C\u002Fsummary>\n\n  **数据与基础设施**\n  - 升级数据库，提升稳定性和性能\n\n  **调度器**\n  - 使用 Redis Streams 和队列隔离重构任务调度器\n  - 新增任务优先级、自动恢复及配额制调度功能\n\n  **部署与工程**\n  - 新增轻量级部署模式，提供快速与完整两种选择\n\n  \u003C\u002Fdetails>\n\n  \u003Cdetails>\n    \u003Csummary>🐞 \u003Cb>Bug 修复\u003C\u002Fb>\u003C\u002Fsummary>\n\n  **记忆调度与更新**\n  - 修复了旧版调度 API，确保记忆隔离正确\n  - 修复了记忆更新日志，使其能正确显示新记忆\n\n  \u003C\u002Fdetails>\n\n- **2025年8月7日** · 🎉 **MemOS v1.0.0（MemCube）发布**\n  首个 MemCube 版本，包含文字游戏演示、LongMemEval 评估、BochaAISearchRetriever 集成、搜索能力提升，以及官方 Playground 正式上线。\n\n  \u003Cdetails>\n    \u003Csummary>✨ \u003Cb>新特性\u003C\u002Fb>\u003C\u002Fsummary>\n\n  **Playground**\n  - 扩展了 Playground 功能及算法性能。\n\n  **MemCube 构建**\n  - 基于 MemCube 小说新增文本游戏演示。\n\n  **扩展评估集**\n  - 新增 LongMemEval 评估结果及脚本。\n\n  \u003C\u002Fdetails>\n\n  \u003Cdetails>\n    \u003Csummary>📈 \u003Cb>改进\u003C\u002Fb>\u003C\u002Fsummary>\n\n  **纯文本记忆**\n  - 集成 Bocha 的互联网搜索功能\n  - 扩展图数据库支持\n  - 为树状结构的纯文本记忆搜索界面增加了上下文理解能力\n\n  \u003C\u002Fdetails>\n\n  \u003Cdetails>\n    \u003Csummary>🐞 \u003Cb>Bug 修复\u003C\u002Fb>\u003C\u002Fsummary>\n\n  **KV 缓存拼接**\n  - 修复了 concat_cache 方法。\n\n  **纯文本记忆**\n  - 修复了图搜索相关问题。\n\n  \u003C\u002Fdetails>\n\n- **2025年7月7日** · 🎉 **MemOS v1.0：Stellar（星河）预览版发布**\n  一款面向 LLM 的 SOTA 记忆操作系统现已开源。\n- **2025年7月4日** · 🎉 **MemOS 论文发布**\n  [MemOS：面向 AI 系统的记忆操作系统](https:\u002F\u002Farxiv.org\u002Fabs\u002F2507.03724) 已在 arXiv 上发布。\n- **2024年7月4日** · 🎉 **Memory3 模型于 WAIC 2024 发布**\n  具有记忆分层架构的 Memory3 模型在 2024 年世界人工智能大会上亮相。\n\n\u003Cbr>\n\n## 🚀 快速入门指南\n\n### ☁️ 1、云端 API（托管）\n#### 获取 API 密钥\n- 在 [MemOS 控制台](https:\u002F\u002Fmemos-dashboard.openmem.net\u002Fcn\u002Fquickstart\u002F?source=landing) 注册\n- 进入 **API 密钥** 页面并复制您的密钥\n\n#### 下一步\n- [MemOS 云端入门](https:\u002F\u002Fmemos-docs.openmem.net\u002Fmemos_cloud\u002Fquick_start\u002F)：几分钟内连接到 MemOS 云端并启用记忆功能。\n- [MemOS 云平台](https:\u002F\u002Fmemos.openmem.net\u002F?from=\u002Fquickstart\u002F)：探索云端控制台、功能及工作流程。\n\n### 🖥️ 2、自托管（本地\u002F私有）\n1. 获取代码仓库。\n    ```bash\n    git clone https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS.git\n    cd MemOS\n    pip install -r .\u002Fdocker\u002Frequirements.txt\n    ```\n2. 配置 `docker\u002F.env.example` 并复制到 `MemOS\u002F.env`\n - `OPENAI_API_KEY`、`MOS_EMBEDDER_API_KEY`、`MEMRADER_API_KEY` 等可通过 [`BaiLian`](https:\u002F\u002Fbailian.console.aliyun.com\u002F?spm=a2c4g.11186623.0.0.2f2165b08fRk4l&tab=api#\u002Fapi) 申请。\n - 将相应配置填入 `MemOS\u002F.env` 文件中。\n - 支持的 LLM 提供商：**OpenAI**、**Azure OpenAI**、**Qwen（DashScope）**、**DeepSeek**、**MiniMax**、**Ollama**、**HuggingFace**、**vLLM**。通过设置 `MOS_CHAT_MODEL_PROVIDER` 来选择后端（例如 `openai`、`qwen`、`deepseek`、`minimax`）。\n3. 启动服务。\n\n- 通过 Docker 启动\n  ###### 提示：请确保已成功安装 Docker Compose，并在执行以下命令前进入 docker 目录（使用 `cd docker`）。\n  ```bash\n  # 进入 docker 目录\n  docker compose up\n  ```\n  ##### 详细步骤请参阅[`Docker 参考`](https:\u002F\u002Fdocs.openmem.net\u002Fopen_source\u002Fgetting_started\u002Frest_api_server\u002F#method-1-docker-use-repository-dependency-package-imagestart-recommended-use)。\n\n- 通过 uvicorn 命令行界面（CLI）启动\n  ###### 提示：请确保 Neo4j 和 Qdrant 已运行后再执行以下命令。\n  ```bash\n  cd src\n  uvicorn memos.api.server_api:app --host 0.0.0.0 --port 8001 --workers 1\n  ```\n  ##### 详细集成步骤请参阅 [`CLI 参考`](https:\u002F\u002Fdocs.openmem.net\u002Fopen_source\u002Fgetting_started\u002Frest_api_server\u002F#method-3client-install-with-CLI)。\n\n\n\n### 自托管基本使用\n  - 添加用户消息\n    ```python\n    import requests\n    import json\n\n    data = {\n        \"user_id\": \"8736b16e-1d20-4163-980b-a5063c3facdc\",\n        \"mem_cube_id\": \"b32d0977-435d-4828-a86f-4f47f8b55bca\",\n        \"messages\": [\n            {\n                \"role\": \"user\",\n                \"content\": \"我喜欢草莓\"\n            }\n        ],\n        \"async_mode\": \"sync\"\n    }\n    headers = {\n        \"Content-Type\": \"application\u002Fjson\"\n    }\n    url = \"http:\u002F\u002Flocalhost:8000\u002Fproduct\u002Fadd\"\n\n    res = requests.post(url=url, headers=headers, data=json.dumps(data))\n    print(f\"result: {res.json()}\")\n    ```\n  - 搜索用户记忆\n    ```python\n    import requests\n    import json\n\n    data = {\n        \"query\": \"我喜欢什么\",\n        \"user_id\": \"8736b16e-1d20-4163-980b-a5063c3facdc\",\n        \"mem_cube_id\": \"b32d0977-435d-4828-a86f-4f47f8b55bca\"\n    }\n    headers = {\n        \"Content-Type\": \"application\u002Fjson\"\n    }\n    url = \"http:\u002F\u002Flocalhost:8000\u002Fproduct\u002Fsearch\"\n\n    res = requests.post(url=url, headers=headers, data=json.dumps(data))\n    print(f\"result: {res.json()}\")\n    ```\n\n\u003Cbr>\n\n## 📚 资源\n\n- **Awesome-AI-Memory**\n  这是一个精心整理的资源库，专注于大型语言模型的记忆与记忆系统相关资料。它系统性地收集了相关的研究论文、框架、工具以及实践见解。该仓库旨在梳理并呈现快速发展的LLM记忆研究领域，连接自然语言处理、信息检索、智能体系统和认知科学等多个研究方向。\n- **开始使用** 👉 [IAAR-Shanghai\u002FAwesome-AI-Memory](https:\u002F\u002Fgithub.com\u002FIAAR-Shanghai\u002FAwesome-AI-Memory)\n- **MemOS Cloud OpenClaw 插件**\n  MemOS Cloud 的官方 OpenClaw 生命周期插件。它会在智能体启动前自动从 MemOS 中调用上下文，并在智能体完成任务后将对话保存回 MemOS。\n- **开始使用** 👉 [MemTensor\u002FMemOS-Cloud-OpenClaw-Plugin](https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS-Cloud-OpenClaw-Plugin)\n\n\u003Cbr>\n\n## 💬 社区与支持\n\n加入我们的社区，提问、分享你的项目，并与其他开发者交流。\n\n- **GitHub Issues**: 在我们的 \u003Ca href=\"https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fissues\" target=\"_blank\">GitHub Issues\u003C\u002Fa> 中报告问题或请求功能。\n- **GitHub Pull Requests**: 通过 \u003Ca href=\"https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpulls\" target=\"_blank\">Pull Requests\u003C\u002Fa> 贡献代码改进。\n- **GitHub Discussions**: 参与我们的 \u003Ca href=\"https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fdiscussions\" target=\"_blank\">GitHub Discussions\u003C\u002Fa>,提出问题或分享想法。\n- **Discord**: 加入我们的 \u003Ca href=\"https:\u002F\u002Fdiscord.gg\u002FTxbx3gebZR\" target=\"_blank\">Discord 服务器\u003C\u002Fa>。\n- **微信**: 扫描二维码加入我们的微信群。\n\n\u003Cdiv align=\"center\">\n  \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FMemTensor_MemOS_readme_423b366f315e.png\" alt=\"二维码\" width=\"300\" \u002F>\n\u003C\u002Fdiv>\n\n\u003Cbr>\n\n## 📜 引用\n\n> [!NOTE]\n> 我们于 **2025年5月28日** 公开发布了简版论文，使其成为最早提出面向LLM的记忆操作系统概念的工作。\n\n如果你在研究中使用了 MemOS，我们非常感谢你引用我们的论文。\n\n```bibtex\n\n@article{li2025memos_long,\n  title={MemOS: A Memory OS for AI System},\n  author={Li, Zhiyu and Song, Shichao and Xi, Chenyang and Wang, Hanyu and Tang, Chen and Niu, Simin and Chen, Ding and Yang, Jiawei and Li, Chunyu and Yu, Qingchen and Zhao, Jihao and Wang, Yezhaohui and Liu, Peng and Lin, Zehao and Wang, Pengyuan and Huo, Jiahao and Chen, Tianyi and Chen, Kai and Li, Kehang and Tao, Zhen and Ren, Junpeng and Lai, Huayi and Wu, Hao and Tang, Bo and Wang, Zhenren and Fan, Zhaoxin and Zhang, Ningyu and Zhang, Linfeng and Yan, Junchi and Yang, Mingchuan and Xu, Tong and Xu, Wei and Chen, Huajun and Wang, Haofeng and Yang, Hongkang and Zhang, Wentao and Xu, Zhi-Qin John and Chen, Siheng and Xiong, Feiyu},\n  journal={arXiv preprint arXiv:2507.03724},\n  year={2025},\n  url={https:\u002F\u002Farxiv.org\u002Fabs\u002F2507.03724}\n}\n\n@article{li2025memos_short,\n  title={MemOS: An Operating System for Memory-Augmented Generation (MAG) in Large Language Models},\n  author={Li, Zhiyu and Song, Shichao and Wang, Hanyu and Niu, Simin and Chen, Ding and Yang, Jiawei and Xi, Chenyang and Lai, Huayi and Zhao, Jihao and Wang, Yezhaohui and others},\n  journal={arXiv preprint arXiv:2505.22101},\n  year={2025},\n  url={https:\u002F\u002Farxiv.org\u002Fabs\u002F2505.22101}\n}\n\n@article{yang2024memory3,\nauthor = {Yang, Hongkang and Zehao, Lin and Wenjin, Wang and Wu, Hao and Zhiyu, Li and Tang, Bo and Wenqiang, Wei and Wang, Jinbo and Zeyun, Tang and Song, Shichao and Xi, Chenyang and Yu, Yu and Kai, Chen and Xiong, Feiyu and Tang, Linpeng and Weinan, E},\ntitle = {Memory$^3$: Language Modeling with Explicit Memory},\njournal = {Journal of Machine Learning},\nyear = {2024},\nvolume = {3},\nnumber = {3},\npages = {300--346},\nissn = {2790-2048},\ndoi = {https:\u002F\u002Fdoi.org\u002F10.4208\u002Fjml.240708},\nurl = {https:\u002F\u002Fglobal-sci.com\u002Farticle\u002F91443\u002Fmemory3-language-modeling-with-explicit-memory}\n}\n```\n\n\u003Cbr>\n\n## 🙌 贡献\n\n我们欢迎社区的贡献！请阅读我们的 [贡献指南](https:\u002F\u002Fmemos-docs.openmem.net\u002Fopen_source\u002Fcontribution\u002Foverview\u002F) 以开始参与。\n\n\u003Cbr>\n\n## 📄 许可证\n\nMemOS 采用 [Apache 2.0 许可证](.\u002FLICENSE) 许可。","# MemOS 快速上手指南\n\nMemOS 是一款专为 LLM 和 AI Agent 设计的记忆操作系统，支持长期记忆的存储、检索与管理，具备多模态记忆、知识库管理及企业级优化能力。\n\n## 1. 环境准备\n\n### 系统要求\n- **操作系统**: Linux, macOS, Windows\n- **Python 版本**: 3.8+ (推荐 3.10+)\n- **容器化部署 (推荐)**: Docker & Docker Compose\n- **本地依赖服务** (若不使用 Docker):\n  - Neo4j (图数据库)\n  - Qdrant (向量数据库)\n\n### 前置依赖\n- 确保已安装 `git`, `pip` (或 `uv`), `docker-compose`。\n- **API Key 准备**:\n  - 需准备大模型及 Embedding 服务的 API Key。\n  - 国内用户推荐使用阿里云百炼 (`BaiLian`) 获取通义千问 (Qwen) 等相关密钥。\n  - 支持提供商：OpenAI, Azure OpenAI, Qwen (DashScope), DeepSeek, MiniMax, Ollama, HuggingFace, vLLM。\n\n## 2. 安装步骤\n\n### 方式一：Docker 部署（推荐）\n此方式自动处理所有依赖服务（Neo4j, Qdrant 等），最适合快速启动。\n\n1. **克隆仓库**\n   ```bash\n   git clone https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS.git\n   cd MemOS\n   pip install -r .\u002Fdocker\u002Frequirements.txt\n   ```\n\n2. **配置环境变量**\n   复制示例配置文件并编辑：\n   ```bash\n   cp docker\u002F.env.example MemOS\u002F.env\n   ```\n   编辑 `MemOS\u002F.env`，填入以下关键配置：\n   - `OPENAI_API_KEY` \u002F `MOS_EMBEDDER_API_KEY` \u002F `MEMRADER_API_KEY`: 填入从 [阿里云百炼](https:\u002F\u002Fbailian.console.aliyun.com\u002F) 或其他提供商获取的密钥。\n   - `MOS_CHAT_MODEL_PROVIDER`: 设置后端模型提供商 (例如：`qwen`, `deepseek`, `openai`, `ollama`)。\n\n3. **启动服务**\n   进入 docker 目录并启动：\n   ```bash\n   cd docker\n   docker compose up\n   ```\n\n### 方式二：命令行 (CLI) 部署\n适用于已有 Neo4j 和 Qdrant 运行环境的开发者。\n\n1. **安装依赖并配置环境变量**\n   参考上述“方式一”的步骤 1 和 2 完成代码克隆与 `.env` 配置。\n\n2. **启动服务**\n   确保 Neo4j 和 Qdrant 正在运行后，执行：\n   ```bash\n   cd src\n   uvicorn memos.api.server_api:app --host 0.0.0.0 --port 8001 --workers 1\n   ```\n\n## 3. 基本使用\n\n服务启动后（默认端口 `8000` 或 `8001`），可通过 HTTP API 进行记忆操作。\n\n### 添加用户记忆\n将用户对话内容存入记忆系统：\n\n```python\nimport requests\nimport json\n\ndata = {\n    \"user_id\": \"8736b16e-1d20-4163-980b-a5063c3facdc\",\n    \"mem_cube_id\": \"b32d0977-435d-4828-a86f-4f47f8b55bca\",\n    \"messages\": [\n        {\n            \"role\": \"user\",\n            \"content\": \"I like strawberry\"\n        }\n    ],\n    \"async_mode\": \"sync\"\n}\nheaders = {\n    \"Content-Type\": \"application\u002Fjson\"\n}\n# 根据实际启动端口调整 URL (Docker 默认为 8000, CLI 默认为 8001)\nurl = \"http:\u002F\u002Flocalhost:8000\u002Fproduct\u002Fadd\"\n\nres = requests.post(url=url, headers=headers, data=json.dumps(data))\nprint(f\"result: {res.json()}\")\n```\n\n### 检索用户记忆\n根据查询内容检索相关记忆：\n\n```python\nimport requests\nimport json\n\ndata = {\n    \"query\": \"What do I like\",\n    \"user_id\": \"8736b16e-1d20-4163-980b-a5063c3facdc\",\n    \"mem_cube_id\": \"b32d0977-435d-4828-a86f-4f47f8b55bca\",\n    \"top_k\": 5\n}\nheaders = {\n    \"Content-Type\": \"application\u002Fjson\"\n}\nurl = \"http:\u002F\u002Flocalhost:8000\u002Fproduct\u002Fsearch\"\n\nres = requests.post(url=url, headers=headers, data=json.dumps(data))\nprint(f\"search result: {res.json()}\")\n```\n\n> **提示**: 更多高级功能（如多模态记忆、知识库管理、反馈修正）请参考官方文档或访问 MemOS Dashboard。","一位独立开发者正在构建一个基于 OpenClaw 的多任务自动化助手集群，旨在处理从数据清洗到代码生成的复杂工作流。\n\n### 没有 MemOS 时\n- **技能无法复用**：每个新任务（如“修复 Python 报错”或“优化 SQL 查询”）都需要重新向模型描述操作步骤，导致大量重复的 Prompt 输入。\n- **上下文爆炸与成本高昂**：为了保持连贯性，不得不将完整的聊天历史作为上下文传入，Token 消耗巨大且响应速度随对话轮数增加而显著变慢。\n- **智能体间存在壁垒**：负责数据分析的 Agent 无法利用负责代码编写的 Agent 之前学到的经验，多实例间如同孤岛，无法共享用户偏好或特定领域知识。\n- **长期记忆缺失**：一旦会话结束，之前调试成功的解决方案随即丢失，下次遇到类似问题时系统表现得像“失忆”一样从头开始。\n\n### 使用 MemOS 后\n- **技能持久化与进化**：MemOS 自动将成功解决的任务抽象为“技能记忆”，后续遇到类似需求时直接调用并自我升级，无需重复教学。\n- **精准检索降低开销**：通过混合搜索技术仅提取相关记忆片段而非全量历史，相比传统方式节省约 72% 的 Token 用量，大幅提升响应效率。\n- **跨智能体协同共享**：不同功能的 Agent 通过统一的 `user_id` 共享记忆库，数据分析师发现的规律可立即被代码生成器利用，实现无缝协作。\n- **本地\u002F云端灵活部署**：开发者可选择本地 SQLite 存储确保数据隐私，或使用云服务实现多设备间的记忆同步，确保持续的学习能力不中断。\n\nMemOS 通过赋予 AI 系统持久的技能记忆与进化能力，将一次性对话转化为可积累、可复用的智能资产，彻底解决了长程任务中的遗忘与高成本难题。","https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FMemTensor_MemOS_2c4094df.png","MemTensor","OpenMem","https:\u002F\u002Foss.gittoolsai.com\u002Favatars\u002FMemTensor_94b8a217.jpg","",null,"https:\u002F\u002Fgithub.com\u002FMemTensor",[80,84,88,92,96,100,104,108,112,115],{"name":81,"color":82,"percentage":83},"Python","#3572A5",46.1,{"name":85,"color":86,"percentage":87},"TypeScript","#3178c6",45.7,{"name":89,"color":90,"percentage":91},"HTML","#e34c26",5.1,{"name":93,"color":94,"percentage":95},"JavaScript","#f1e05a",2.1,{"name":97,"color":98,"percentage":99},"Shell","#89e051",0.6,{"name":101,"color":102,"percentage":103},"CSS","#663399",0.2,{"name":105,"color":106,"percentage":107},"PowerShell","#012456",0.1,{"name":109,"color":110,"percentage":111},"Dockerfile","#384d54",0,{"name":113,"color":114,"percentage":111},"Makefile","#427819",{"name":116,"color":117,"percentage":111},"Go Template","#00ADD8",8397,739,"2026-04-16T16:05:27","Apache-2.0","Linux, macOS, Windows","未说明（支持本地部署，依赖具体选择的 LLM 后端，如 Ollama 或 vLLM 可能需 GPU）","未说明",{"notes":126,"python":127,"dependencies":128},"本地自托管版本需要预先安装并运行 Neo4j（图数据库）、Qdrant（向量数据库）和 Redis（任务调度）。支持多种大模型后端（OpenAI, Azure, Qwen, DeepSeek, Ollama 等），需配置相应的 API Key。推荐使用 Docker Compose 进行一键部署，也可通过 CLI 手动启动。","3.8+",[129,130,131,132,133],"Neo4j","Qdrant","Redis","uvicorn","requests",[13,14,135,35],"其他",[137,138,139,140,141,142,143,144,145,146,147,148,149,150,151,6,152,153],"agent","llm","memory","llm-memory","long-term-memory","memory-management","memory-retrieval","memory-scheduling","memory-operating-system","rag","retrieval-augmented-generation","agent-memory","memory-agent","skill-memory","skills","clawdbot","moltbot","2026-03-27T02:49:30.150509","2026-04-17T09:46:59.436149",[157,162,167,172,177,182],{"id":158,"question_zh":159,"answer_zh":160,"source_url":161},37152,"为什么输入中文时记忆结果不正确或变成英文？","这是因为 Ollama 中的 nomic-embed-text:latest 模型官方不支持中文。建议更换为支持多语言的 Embedding 后端，例如 OpenAI 的 Embedding API。您可以在 examples\u002Fbasic_modules\u002Fembedder.py 中找到对应的配置示例。","https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fissues\u002F130",{"id":163,"question_zh":164,"answer_zh":165,"source_url":166},37153,"如何配置 MemOS 以支持高并发场景下的 vLLM 后端？","MemOS 最新版本已正式支持 vLLM。您可以将配置文件中的 llm backend 设置为 \"openai\"，并将 api_base 指向 vLLM 的服务地址（如 http:\u002F\u002Fyour-vllm-url\u002Fv1\u002F），api_key 设为任意值（如 \"EMPTY\" 或实际 key）。Embedder 也可类似配置，provider 设为 \"openai\"，base_url 指向 vLLM 的 embedding 接口。具体配置示例如下：\n{\n  \"llm\": {\n    \"backend\": \"openai\",\n    \"config\": {\n      \"api_base\": \"http:\u002F\u002F\u003Cvllm-ip>:\u003Cport>\u002Fv1\u002F\",\n      \"api_key\": \"EMPTY\",\n      \"model_name_or_path\": \"your-model-name\",\n      \"temperature\": 0.0,\n      \"max_tokens\": 8192\n    }\n  },\n  \"embedder\": {\n    \"backend\": \"universal_api\",\n    \"config\": {\n      \"provider\": \"openai\",\n      \"api_key\": \"EMPTY\",\n      \"model_name_or_path\": \"embedding-model\",\n      \"base_url\": \"http:\u002F\u002F\u003Cvllm-ip>:\u003Cport>\u002Fv1\u002F\"\n    }\n  }\n}\n更多细节可参考示例文件：examples\u002Fmem_os\u002Fsimple_vllm_memos.py","https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fissues\u002F24",{"id":168,"question_zh":169,"answer_zh":170,"source_url":171},37154,"初始化 SentenceChunker 时报错 'unexpected keyword argument tokenizer_or_token_counter' 怎么办？","该问题是由于 chonkie 库升级到 1.4.0 版本后引入了破坏性变更（Breaking Change），导致参数不兼容。目前社区已提交 PR（#705）修复此问题。临时解决方案是等待官方发布新版本，或手动回退 chonkie 到 1.4.0 以下版本，直到官方修复合并。","https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fissues\u002F578",{"id":173,"question_zh":174,"answer_zh":175,"source_url":176},37155,"MemOS 是否支持沙箱环境（如 BoxLite）运行？","MemOS 定位为记忆操作系统，主要负责记忆的存储、检索和组织，不负责执行用户代码或运行 Agent 任务，因此官方并未内置沙箱支持。沙箱功能通常属于 Agent 执行层。不过，用户可以在外部容器或微 VM（如 BoxLite）中部署 MemOS 实例，以实现隔离运行，但这需要用户自行集成，不属于 MemOS 核心功能。","https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fissues\u002F1128",{"id":178,"question_zh":179,"answer_zh":180,"source_url":181},37156,"如何实现多角色或多视角的记忆管理（Multi-View Memory）？","MemOS 支持通过 role_id 或 role_name 字段自动检测并切换多视角记忆模式。每个角色拥有独立的记忆空间，同时可共享全局世界记忆。典型流程为：游戏事件写入世界记忆 → NPC 对话时结合世界记忆与角色专属记忆生成回复 → 对话结果写回对应角色的记忆空间。该机制适用于多 Agent 系统、游戏 NPC 等场景，具体实现可参考相关增强提案及示例代码。","https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fissues\u002F1084",{"id":183,"question_zh":184,"answer_zh":185,"source_url":161},37157,"在哪里可以找到 Embedder 模块的配置示例？","Embedder 模块的配置示例位于项目目录下的 examples\u002Fbasic_modules\u002Fembedder.py 文件中。该文件展示了如何配置不同后端（如 OpenAI、Ollama 等）的 Embedding 服务，包括 API 地址、密钥、模型名称等关键参数，推荐用户参考此文件进行自定义配置。",[187,192,197,202,207,212,217,222,227,232,237,242,247,252,257,262,267,272,277,282],{"id":188,"version":189,"summary_zh":190,"released_at":191},297658,"v2.0.13","## 变更内容\n* chore：将内存迁移回系统提示，由 @hijzy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1330 中完成\n* fix(memos-local)：按 hubInstanceId 分割共享状态，并修复所有者问题……由 @tangbotony 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1331 中完成\n* feat(memos-local-openclaw)：引入 Hub 嵌入与并行召回功能，修复自动更新可靠性问题，由 @tangbotony 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1341 中完成\n* fix：安装包时使用 npm 替代 npx，由 @hijzy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1335 中完成\n* fix(memos-local-openclaw)：解决 OpenClaw fallb 中 SecretRef apiKey 的解析问题……由 @hijzy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1345 中完成\n* feat：修改内存共享方式，由 @Wang-Daoji 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1368 中完成\n* feat：为本地插件注册上下文引擎槽位，由 @hijzy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1356 中完成\n* fix：更新云插件代码，由 @Hun-ger 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1385 中完成\n* Openclaw 本地插件 20260324，由 @syzsunshine219 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1386 中完成\n* fix(openclaw)：去重 postinstall 重建块（SyntaxError: startMs… 已声明），由 @syzsunshine219 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1388 中完成\n* add MemOS 插件一键安装技能，由 @Mathematics-Yang 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1394 中完成\n* 更新本地插件代码，由 @Hun-ger 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1396 中完成\n* 更新本地插件代码，由 @Hun-ger 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1397 中完成\n* fix：提升 better-sqlite3 对 Node.js v25 的兼容性，由 @salws 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1212 中完成\n* fix(内存)：规范化 \u002Fnew 启动提示的自动召回查询，由 @Naluko 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1228 中完成\n* fix：当存在通配符 '*' 时，跳过 tools.allow 的打补丁操作，由 @jamesMuWB 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1393 中完成\n* fix(安装)：处理 postinstall 中的 node_modules 清理工作，并移除过时的 contextEngine 槽位，由 @hijzy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1399 中完成\n* fix(安装)：在 openclaw.json 的 slots、entries 和 installs 中注册插件，由 @hijzy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1402 中完成\n* fix(安装)：改用网关服务而非前台运行，并添加完成提示，由 @hijzy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1406 中完成\n* feat(查看器)：为任务\u002F技能添加会话筛选和批量删除功能，由 @Hun-ger 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1407 中完成\n* feat(memos-local)：结构化主题分类器、智能体追踪以及任务自动完结功能，由 @tangbotony 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1412 中完成\n* fix(openclaw-plugin)：自启动查看器及安装脚本优化，由 @hijzy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1413 中完成\n* fix(查看器)：修复并发选择箭头重叠问题，并限制 taskAutoFinalize 的输入，由 @hijzy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1414 中完成\n* feat(查看器)：添加所有者徽章、统一会话密钥格式，并改进会话显示效果，由 @tangbotony 在 https:\u002F\u002Fgithub.com\u002FMemTenso 中完成","2026-04-10T07:50:50",{"id":193,"version":194,"summary_zh":195,"released_at":196},297659,"v2.0.12","## 变更内容\n* chore: 恢复将内存移回系统提示的操作，由 @CaralHsi 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1330 中完成\n* fix(memos-local): 通过 hubInstanceId 实现状态共享，并修复所有者问题……由 @tangbotony 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1331 中完成\n* feat: 修改内存共享功能，由 @Wang Daoji 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1368 中完成\n\n## 新贡献者\n\n**完整变更日志**: https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fcompare\u002Fv2.0.11...v2.0.12","2026-04-07T03:28:30",{"id":198,"version":199,"summary_zh":200,"released_at":201},297660,"v2.0.11","## 变更内容\n* 修复：@hijzy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1295 中修复了 Windows 目录错误\n* 杂项：@hijzy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1299 中更新了网站文档\n* 功能（memos-local-openclaw）：@tangbotony 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1321 中添加了多 OpenClaw 共享、双实例隔离以及查看器改进功能\n* 功能：@wustzdy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1326 中优化了线程池\n* 功能：@hijzy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1323 中添加了安装脚本\n* 修复：@hijzy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1322 中添加了 DeepSeek 提供者\n* 修复：@duoyidavid-eng 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1312 中将自动召回逻辑从 before_agent_start 迁移到 before_prompt_build\n* 修复：@duoyidavid-eng 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1310 中禁用了智谱思考模式，以防止摘要生成器输出被截断\n* 修复（capture）：@hijzy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1302 中过滤了系统注入的启动\u002F心跳\u002F哨兵回复信息\n* 修复：@CaralHsi 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1309 中正确报告了导入步骤失败的情况\n* 撤销：“功能：添加安装脚本”——@hijzy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1329 中完成\n* Openclaw 本地插件 20260319 —— @syzsunshine219 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1328 中完成\n* 功能：@wustzdy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1337 中优化了针对 PolarDB 的 export_graph 功能\n* 重构：@CaralHsi 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1344 中整合 API 至 server_api，并移除了已弃用的接口\n* 修复：@CaralHsi 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1350 中通过请求上下文改进了 API 验证错误信息\n* 功能：@octo-patch 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1291 中将 MiniMax 默认模型升级至 M2.7\n* 修复：@brentkearney 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1275 中在 MCP add_memory 工具中以聊天格式发送消息\n* 添加用于服务状态监控的健康检查端点 —— @tan-huynh 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1248 中完成\n* 紧急修复：反馈归档新节点 —— @whipser030 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1243 中完成\n* 功能：@wustzdy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1363 中优化了池\n* 修复（API）：支持按 user_id\u002Fconversation_id 删除记忆（修复 #1103）—— @fancyboi999 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1104 中完成\n* 修复：@CaralHsi 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1359 中对 Neo4j 向量搜索进行了预过滤和元数据处理\n* 功能：@Tavily-FDE-Bot 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1357 中添加了 Tavily 作为可配置的互联网搜索后端\n* 修复（graph_dbs）：@kyan-du 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1178 中对嵌套元数据进行清理后再由 Neo4j 写入\n* 功能：@ombmh 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1293 中添加了用于 Kubernetes 部署的 Dockerfile 和 Helm Chart\n* 格式化 —— @CaralHsi 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1369 中完成\n* 开发 2026032","2026-03-27T09:31:57",{"id":203,"version":204,"summary_zh":205,"released_at":206},297661,"v2.0.10","## 变更内容\n* 功能（memreader\u002FLLM）：由 @fridayL 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1246 中添加了 OpenAI memreader 的备份配置\n* 功能（Scheduler）：由 @fridayL 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1252 中将 Scheduler 的默认 LLM 更改为 general_llm\n* 修复：文件名也被存储为记忆的一部分的问题，由 @whipser030 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1222 中修复\n* 修复（graph_dbs）：向 Neo4jCommunityGraphDB 添加缺失的 status 参数，由 @baranchen 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1179 中完成\n* 功能（memcube）：单个 cube 使用 general_llm，由 @fridayL 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1253 中实现\n* 修复：修复计划任务未使用 Redis 的问题，由 @wustzdy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1267 中完成\n* 功能：添加测试覆盖率 CI，由 @CarltonXiang 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1272 中实现\n* 功能：添加带有预构建的 openclaw 插件发布工作流，由 @CarltonXiang 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1278 中完成\n* 功能：从 native_memos 同步 memos-local-openclaw，由 @syzsunshine219 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1279 中完成\n* 功能：优化 search_by_embedding 和日志记录，由 @wustzdy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1284 中实现\n* 修复（openclaw）：从助手消息中移除 \u003Cthink\u002F> 标签，由 @hijzy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1287 中完成\n* 修复（memos-local-openclaw）：在解析 better-sqlite3 时规范化 Windows 路径，由 @hijzy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1265 中完成\n* 修复（memos-local-openclaw）：在查看器迁移中尊重 OPENCLAW_STATE_DIR，由 @hijzy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1262 中完成\n* 修复（memos-local-openclaw）：在 OpenClaw 回退配置中检测 Anthropic 提供商，由 @hijzy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1263 中完成\n* 功能：修复 2.0.10 版本的 bug，由 @Wang-Daoji 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1282 中实现\n* 修复（memos-local-openclaw）：在回退模型配置中尊重 OPENCLAW_STATE_DIR，由 @hijzy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1264 中完成\n* 修复（OpenClaw 本地插件）：解决搜索和 SQLite 安装问题，由 @tangbotony 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1251 中完成\n* 检测 allowPromptInjection 并优化自动召回处理，由 @lcpdeb 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1285 中完成\n* 功能（memreader）：更新 prefermem 配置，由 @fridayL 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1294 中完成\n* 修复（openclaw）：检测 Anthropic 提供商并使用正确的 API 格式，由 @zerone0x 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1260 中完成\n* 功能（memos-local-openclaw）：此 PR 合并了本地插件 0317 分支（openclaw-local-plugin-20260317），由 @syzsunshine219 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1289 中完成\n* 添加 MemOS-Cloud-OpenClaw-Plugin 的 Git Subtree，由 @XiaohuiSu 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1297 中完成\n* 开发 20260319 v2.0.10，由 @CaralHsi 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1255 中完成\n\n## 新贡献者\n* @baranchen 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1179 中做出了首次贡献\n* @lcpdeb 做出了其","2026-03-20T09:48:31",{"id":208,"version":209,"summary_zh":210,"released_at":211},297662,"v2.0.9","## 变更内容\n* 功能：添加 auto_cleanup_working；由 @CaralHsi 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1169 中添加 image_id\u002Ffile_info\n* 修复\u002F返回消息：由 @CaralHsi 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1175 中完成\n* 功能\u002F合并 main 2.0.9：由 @CaralHsi 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1190 中完成\n* 功能（memreader）：添加通用配置及训练 memreader 模型，由 @fridayL 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1185 中完成\n* 功能（memreader）：将图像 LLM 解析器回调至通用 LLM，由 @fridayL 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1195 中完成\n* 功能：优化 export_graph，由 @wustzdy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1197 中完成\n* 功能（memreader）：将文件内容中的图像 LLM 设置为默认模型，由 @fridayL 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1198 中完成\n* 功能：添加 openclaw-local-plugin 版本（memos-local-openclaw），由 @syzsunshine219 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1205 中完成\n* 修复：project_id\u002Fmanager_id 问题，由 @CaralHsi 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1208 中完成\n* 功能：优化 pool && log，由 @wustzdy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1210 中完成\n* 修复：在建议处理器中处理字符串类型消息，由 @CaralHsi 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1214 中完成\n* 开发 20260309 v2.0.9：由 @CaralHsi 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1238 中完成\n\n## 新贡献者\n* @syzsunshine219 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1205 中完成了首次贡献\n\n**完整变更日志**：https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fcompare\u002Fv2.0.8...v2.0.9","2026-03-16T07:33:42",{"id":213,"version":214,"summary_zh":215,"released_at":216},297663,"v2.0.8","## 变更内容\n* 修复：将 AuthConfig 部分初始化的日志级别由 WARNING 降级为 INFO，由 @anatolykoptev 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1052 中完成。\n* 修复：修复激活内存配置在调用 OpenAI 的 get_default() 方法时导致的崩溃问题，由 @anatolykoptev 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1051 中完成。\n* 修复：为云端嵌入器添加 Unicode 校验功能，由 @anatolykoptev 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1048 中完成。\n* 杂项：从 main 分支同步代码，由 @CaralHsi 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1147 中完成。\n* 新特性：优化知识库的分块逻辑，由 @whipser030 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1135 中完成。\n* 新特性：将偏好设置迁移至 PolarDB 数据库，由 @Wang-Daoji 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1145 中完成。\n* 新特性：优化 PolarDB 的线程池连接池，由 @wustzdy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1152 中完成。\n* 修复：在偏好记忆中使用相对性而非分数，由 @hijzy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1153 中完成。\n* 修复：修复图像相关 bug；多模态阅读器中单个项目没有嵌入向量的问题，由 @CaralHsi 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1154 中完成。\n* 修复：当 add_nodes_batch 中缺少嵌入向量时不再抛出错误，由 @bittergreen 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1157 中完成。\n* 新特性：使通用\u002F朴素记忆与树形记忆对齐，由 @CaralHsi 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1159 中完成。\n* 新特性：更新日志记录，由 @wustzdy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1161 中完成。\n* 新特性：优化全文搜索功能，由 @wustzdy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1164 中完成。\n* 修复：删除无用的图召回功能，由 @hijzy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1167 中完成。\n* 新特性：再次优化全文搜索功能，由 @wustzdy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1170 中完成。\n* 新特性：进一步优化全文搜索功能，由 @wustzdy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1171 中完成。\n* 修复：由于未能传递 'user_name' 参数，反馈功能在调用搜索接口时会失败，由 @whipser030 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1174 中完成。\n* 新特性：优化配置，由 @wustzdy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1176 中完成。\n* 新特性：为 OpenClaw 添加持久化本地记忆插件（memos-local-openclaw），由 @tangbotony 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1182 中完成。\n* 文档：更新 README 文件，并将 memos-local-openclaw 插件版本升级至 v1.0.0，由 @tangbotony 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1184 中完成。\n* 杂项：将版本号修改为 v2.0.8，由 @wustzdy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1186 中完成。\n* 开发版 20260302 v2.0.8，由 @CaralHsi 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1148 中完成。\n\n\n**完整变更日志**：https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fcompare\u002Fv2.0.7...v2.0.8","2026-03-09T10:01:54",{"id":218,"version":219,"summary_zh":220,"released_at":221},297664,"v2.0.7","## 变更内容\n* 修复：由 @hijzy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1095 中添加了针对 Neo4j 数据库的全文搜索功能。\n* 修复：由 @hijzy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1096 中增加了全文检索路径的切换开关。\n* 修复：由 @CaralHsi 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1098 中为全文检索路径切换开关（#1096）进行了补充。\n* 修复：由 @CaralHsi 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1099 中为 Neo4j 数据库添加了全文搜索功能（#1095）。\n* 功能：由 @lailoo 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1097 中向搜索方法添加了 return_fields 参数（#955）。\n* 功能：由 @wustzdy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1111 中优化了 user_name、关键词及日志记录。\n* 紧急修复：由 @whipser030 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1109 中修复了在添加图边时出现的错误。\n* 修复：由 @glin93 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1116 中修复了基于 user_name 等范围的新增\u002F更新日志判断逻辑。\n* 修复：由 @hijzy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1119 中设置了默认的相关性。\n* 修复：由 @wustzdy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1123 中移除了未使用的边，并优化了日志记录。\n* 修复：由 @whipser030 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1125 中将文件内存解析结果输出为列表类型。\n* 功能：由 @CaralHsi 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1126 中从主分支同步代码。\n* 修复：由 @glin93 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1118 中使用 merged_from 正确识别新增\u002F更新的内存日志。\n* 修复：由 @bittergreen 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1130 中修复了 memory_type 验证错误。\n* 修复：由 @CaralHsi 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1131 中处理了 user_name 相关问题。\n* 修复：由 @hijzy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1133 中改为从数据库获取预设内存的嵌入向量，而非重新计算。\n* 修复：由 @haosenwang1018 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1112 中将 thread_safe_dict_segment 中的 bare except 替换为 except Exception。\n* 文档：由 @eltociear 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F898 中更新了 README.md 文件。\n* 功能：由 @wustzdy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1138 中优化了 get_edges 方法。\n* 杂项：由 @bittergreen 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1140 中将版本号修改为 2.0.7。\n* 开发 20260224 v2.0.7：由 @CaralHsi 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1127 中完成。\n\n## 新贡献者\n* @lailoo 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1097 中做出了首次贡献。\n* @haosenwang1018 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1112 中做出了首次贡献。\n* @eltociear 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F898 中做出了首次贡献。\n\n**完整变更日志**：https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fcompare\u002Fv2.0.6...v2.0.7","2026-02-28T09:31:47",{"id":223,"version":224,"summary_zh":225,"released_at":226},297665,"v2.0.6","## 变更内容\n* 修复：允许 `get_user_names_by_memory_ids` 支持所有类型的数据库，由 @tangg555 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1063 中实现。\n* 📃 文档：更新 README 文件，由 @XiaohuiSu 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1070 中完成。\n* 新特性：v2.0.6 版本优化，由 @Wang-Daoji 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1071 中实现。\n* 新特性：添加关键词搜索功能，由 @hijzy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1073 中实现。\n* 修复：将 `pref_mem` 的相关性替换为分数，由 @hijzy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1077 中完成。\n* 修复：支持同时存在字段条件和逻辑运算符的情况，由 @endxxxx 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1076 中完成。\n* 修复：LLM 调用启用思考模式时的 bug，由 @Wang-Daoji 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1080 中完成。\n* 修复：在筛选字段中支持小时、分钟和秒格式，由 @endxxxx 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1081 中完成。\n* 修复：LLM 调用启用思考模式时的 bug，由 @Wang-Daoji 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1082 中完成。\n* 修复：LLM 调用启用思考模式时的 bug，由 @Wang-Daoji 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1083 中完成。\n* 修复：偏好返回结果的问题，由 @CaralHsi 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1088 中完成。\n* 新特性：v2.0.6 版本优化，由 @Wang-Daoji 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1090 中实现。\n* 开发 20260209 v2.0.6，由 @Wang-Daoji 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1091 中完成。\n\n\n**完整变更日志**：https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fcompare\u002Fv2.0.5...v2.0.6","2026-02-12T11:52:32",{"id":228,"version":229,"summary_zh":230,"released_at":231},297666,"v2.0.5","## 变更内容\n* 新增功能：由 @bittergreen 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F992 中初始化用于管理内存版本的数据结构和类。\n* 修复：由 @whipser030 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F995 中避免将 fileurl 添加到 memoryvalue。\n* 新增功能：由 @wustzdy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1001 中添加 delete_node_by_mem_cube_id 和 recover_memory_by_mem_kube_id 函数。\n* 新增功能：由 @anatolykoptev 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F966 中添加 PostgreSQL + pgvector 图数据库后端。\n* 修复：由 @whipser030 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F999 中修复“tags”缺少与内存内容相关关键词的问题。\n* 修复：由 @wustzdy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1007 中添加 ruff 代码检查及 poetry run ruff format 命令。\n* 修复：由 @endxxxx 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1011 中并行处理大内存项的分块，并对这些分块进行处理，但……\n* 修复：由 @wustzdy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1013 中修复 polardb 参数错误。\n* 新增功能：由 @Wang-Daoji 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1016 中优化分支 2.0.5。\n* 修复：由 @whipser030 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1018 中修复“获取内存”接口返回的“节点总数”信息不准确的问题。\n* 文档更新：由 @XiaohuiSu 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1023 中将 MemOS Cloud OpenClaw 插件添加到资源列表。\n* 新增功能：由 @wustzdy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1028 中添加 neo4j 和 neo4j_community 功能。\n* 新增功能：由 @hijzy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1006 中添加相关性阈值过滤及检索内存的最大限制。\n* 新增功能：由 @Wang-Daoji 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1031 中实现技能记忆功能。\n* 重构：由 @fancyboi999 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1004 中对调度器的处理器和搜索管道进行模块化改造。\n* 修复：由 @hijzy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1036 中设置默认相关性。\n* 修复：由 @whipser030 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1037 中再次修复“获取内存”接口返回的“节点总数”信息不准确的问题。\n* 新增功能：由 @wustzdy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1040 中更改代码结构并更新 neo4j。\n* 新增功能：由 @anpulin 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1034 中修改 SDK。\n* 修复：由 @CarltonXiang 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1044 中修复格式错误。\n* 紧急修复：由 @whipser030 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1055 中恢复代码，并用原始来源替换知识库中的内容。\n* 新增功能：由 @Wang-Daoji 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1042 中实现技能记忆功能。\n* 修复：由 @CarltonXiang 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1046 中修复格式错误。\n* 新增功能：由 @Wang-Daoji 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1056 中实现技能记忆功能。\n* 新增功能：由 @glin93 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F1058 中实现 mem read rawfile 处理功能。\n* 新增功能：由 @tangg555 在 https:\u002F\u002Fgithub.com\u002FMemTens 中实现召回知识库内容的原文，而非检索生成的内存。","2026-02-10T02:51:17",{"id":233,"version":234,"summary_zh":235,"released_at":236},297667,"v2.0.4","## 变更内容\n* 功能：@CaralHsi 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F918 中添加镂空功能\n* 合并：@hijzy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F947 中将 dev v2.0.3 合并到 main 分支\n* 功能：@bittergreen 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F958 中为快速添加流程添加记忆提醒功能\n* 杂项：@CarltonXiang 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F959 中更新了 PR 模板\n* 功能：@Wang-Daoji 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F964 中实现技能记忆功能\n* 修复：@CaralHsi 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F968 中修复了游乐场聊天中的 bug\n* 功能：@wustzdy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F971 中修复了 get_nodes 方法\n* 功能：@Wang-Daoji 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F972 中再次实现技能记忆功能\n* 功能：@Wang-Daoji 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F974 中继续实现技能记忆功能\n* 修复：@CarltonXiang 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F975 中修复了游乐场聊天中的 bug\n* 修复：@CaralHsi 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F973 中修复了游乐场聊天中的 bug，并改进了添加流程中的 search_by_embedding 功能\n* 功能：@bittergreen 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F970 中添加了用于快速检测重复或冲突的 NLI 模型\n* 功能：@Wang-Daoji 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F981 中实现技能记忆功能\n* 功能：@hijzy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F954 中添加了 MMR 去重功能\n* 功能：@Wang-Daoji 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F982 中再次实现技能记忆功能\n* 修复：@endxxxx 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F983 中修复了嵌入服务偶尔出现的超时问题\n* 功能：@Wang-Daoji 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F984 中实现技能记忆功能\n* 功能：@CaralHsi 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F988 中调整了技能提示，并修复了一些 bug\n* 修复：@CaralHsi 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F990 中修复了技能线程相关的 bug\n* Dev 20260126 v2.0.4：@CarltonXiang 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F976 中完成\n\n## 新贡献者\n* @hijzy 在 https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F947 中完成了首次贡献\n\n**完整变更日志**：https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fcompare\u002Fv2.0.3...v2.0.4","2026-01-30T09:09:21",{"id":238,"version":239,"summary_zh":240,"released_at":241},297668,"v2.0.3","## What's Changed\r\n* feat: refactor & reorganize examples with unified structure and updated demos by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F903\r\n* fix: refine config for rerank input is too long by @whipser030 in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F920\r\n* fix: modifying api examples by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F917\r\n* fix: component_init initialize scheduler every time by @tangg555 in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F915\r\n* fix: parse in sim-struct chat-data; unify image analysis prompt with context injection & improve LLM robustness by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F913\r\n* fix: json loads error by @whipser030 in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F937\r\n* feat: sync main by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F944\r\n* feat: model error traceback by @CarltonXiang in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F942\r\n* feat: delete_time && delete_record_id for neo4j,polardb,neo4j_community by @wustzdy in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F945\r\n\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fcompare\u002Fv2.0.2...v2.0.3","2026-01-23T06:38:06",{"id":243,"version":244,"summary_zh":245,"released_at":246},297669,"v2.0.2","## What's Changed\r\n* fix: knowledge base adopt raw text by @whipser030 in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F836\r\n* Feat\u002Foptimize cloud service api by @Wang-Daoji in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F839\r\n* Feat\u002Ffix palyground bug by @Wang-Daoji in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F841\r\n* refactor & fix bugs: fix a range of bugs in scheduler and revise fine add apis by @tangg555 in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F840\r\n* feat: merge dev 0112 by @CarltonXiang in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F854\r\n* feat: delete result by @CarltonXiang in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F852\r\n* fix: unuse rerank by @whipser030 in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F855\r\n* fix: backtrack knowledge retrieval by @whipser030 in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F857\r\n* Feat\u002Ftool mem related by @Wang-Daoji in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F864\r\n* docs: fix server start cmd by @Nyakult in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F858\r\n* fix: Qdrant empty when using neo4j-community by @OhhhhPi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F843\r\n* fix: use knowledge embedding score rerank by @whipser030 in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F867\r\n* fix: optimization_and_fix by @Wang-Daoji in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F868\r\n* Dev 20260112 v2.0.2 by @CarltonXiang in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F877\r\n* feat: merge dev into main; add scheduler examples, API enhancements, and retrieval dedup by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F775\r\n* chore: update version number to v2.0.2 by @CarltonXiang in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F879\r\n* feat: prevent repeated memories via archival and activated-only retrieval by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F880\r\n* feat: optimization and fix branch by @Wang-Daoji in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F882\r\n* feat: LLM\u002Fplugin retrieval bug,  prevent redundant memories via archival with activated-only retrieval, and update implicit preference logic by @endxxxx in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F893\r\n\r\n## New Contributors\r\n* @OhhhhPi made their first contribution in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F843","2026-01-16T08:56:28",{"id":248,"version":249,"summary_zh":250,"released_at":251},297670,"v2.0.1","## What's Changed\r\n* update requirements by @pursues in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F772\r\n* update scheduler and add operation for dehallucination by @tangg555 in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F769\r\n* fix: update README.md by @zZhangSir in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F774\r\n* feat： update readme by @zZhangSir in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F776\r\n* feat: add export_graph data page by @wustzdy in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F778\r\n* fix: optimize Neo4j Community Edition support and enhance MCP environment loading by @fancyboi999 in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F754\r\n* fix: add feedback change to preference by @whipser030 in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F771\r\n* fix: improve chat playground stability and chat handler initialization by @Wang-Daoji in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F770\r\n* fix: update README.md by @zZhangSir in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F786\r\n* update: requirement,Dockerfile by @pursues in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F784\r\n* Feat: add OpenAI log by @CarltonXiang in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F785\r\n* Feat\u002Fdedup playground display by @Wang-Daoji in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F789\r\n* hotfix: redis dependency in scheduler by @tangg555 in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F781\r\n* add get_user_names_by_memory_ids api by @Wang-Daoji in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F790\r\n* feat: add batch delete by @wustzdy in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F787\r\n* feat: add timer log by @CarltonXiang in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F793\r\n* feat: add dedup search param by @glin93 in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F788\r\n* Dev zdy 1226 page by @wustzdy in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F796\r\n* requirements add neo4j  by @pursues in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F792\r\n* Main zhq readme by @zZhangSir in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F794\r\n* update docker.compose.yml by @pursues in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F798\r\n* feat: fix requirements by @lijicode in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F800\r\n* Patch: get_memory adds the page size parameter function and the filtering of user id by @whipser030 in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F801\r\n* Dev zdy 1229 by @wustzdy in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F802\r\n* feat: add get_user_names_by_memory_ids for polardb && neo4j by @wustzdy in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F803\r\n* feat: add export_graph total by @wustzdy in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F804\r\n* add: get_memory return edges and count of items by @whipser030 in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F805\r\n* Scheduler: address some issues to run old scheduler example and kv cache example by @tangg555 in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F797\r\n* fix: issues caused by no reading default use_redis from env by @tangg555 in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F799\r\n* add: milvus return data pagination by @whipser030 in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F806\r\n* feat: update source return and  chunk settings by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F808\r\n* add neo4j by @pursues in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F809\r\n* Scheduler: update exampels by @tangg555 in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F807\r\n* feat: add delete_node_by_prams filter by @wustzdy in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F810\r\n* fix: merge all preference by @whipser030 in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F811\r\n* feat: update _build_filter_conditions_sql in conditions && build_cypher_filter_condition filter by @wustzdy in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F812\r\n* feat: _build_filter_conditions_sql filter by @wustzdy in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F813\r\n* fix: update deprecated APIs for chonkie v1.4.0 and qdrant-client v1.16.0 by @zhixiangxue in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F705\r\n* feat: update code format by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F814\r\n* Feat\u002Foptimize cloud service api by @Wang-Daoji in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F816\r\n* fix: [PrefEval Evaluation] propagate --lib and --version arguments in search and response modes by @fancyboi999 in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F780\r\n* fix: fix context error and empty embedding by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F817\r\n* Feat\u002Foptimize cloud service api by @Wang-Daoji in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F818\r\n* fix: logs context and empty embedding by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F819\r\n* Feat\u002Foptimize cloud service api by @Wang-Daoji in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F820\r\n* add exist_user_name for neo4j.py by @wustzdy in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F821\r\n* Feat\u002Foptimize cloud service api by @Wang-Daoji in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F822\r\n* fix reranker by @pursues in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F823\r\n* Feat\u002Foptimize cloud service api by @Wang-Daoji in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fp","2026-01-09T11:20:32",{"id":253,"version":254,"summary_zh":255,"released_at":256},297671,"v2.0.0","## What's Changed\r\n* feat: simplify simple tree by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F461\r\n* feat: max worker by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F475\r\n* Feat\u002Fdedup mem by @Wang-Daoji in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F473\r\n* feat: add topk for working mem by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F476\r\n* feat: update readme by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F477\r\n* feat: change url by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F478\r\n* fix: chat time issue 2023->2025 by @whipser030 in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F479\r\n* scheduler feat: implementation of redis queue and new api search functions of mixture and fine mode by @tangg555 in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F462\r\n* feat: update readme by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F481\r\n* add pool health && log by @wustzdy in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F482\r\n* Feat: reorganize playground code and merge to API  by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F488\r\n* Fix\u002Fno response by @CarltonXiang in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F490\r\n* fix lack mem_cube_id bug in pref async by @Wang-Daoji in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F494\r\n* Feat\u002Ffix explicit threshold by @Wang-Daoji in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F495\r\n* Feat：reorg playground code by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F497\r\n* Feat: add redis_scheduler by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F499\r\n* Update API Reference link in README.md by @CarltonXiang in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F500\r\n* hotfix bug in pref init by @Wang-Daoji in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F502\r\n* feat & refactor: add searcher to handler_init and remove logger info from get_messages by @tangg555 in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F501\r\n* feat: log support time rotating by @CarltonXiang in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F504\r\n* fix init bug pref by @Wang-Daoji in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F508\r\n* Feat\u002Flog rotating by @CarltonXiang in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F507\r\n* fix: format error by @CarltonXiang in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F510\r\n* Dev by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F496\r\n* Feature\u002Fplayground memcube log structured logs by @glin93 in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F509\r\n* feat: sync  change by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F512\r\n* hotfix:hotfix by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F513\r\n* feat: abstract CubeView to Add & Search Handler by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F498\r\n* Feat\u002Fmerge api refactor to dev by @Wang-Daoji in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F514\r\n* Feature\u002Fmemcube log structured logs rework by @glin93 in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F516\r\n* feat: enhance APIADDRequest with custom_tags, info, and is_feedback fields by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F515\r\n* feat: add headers for embedding and reranker by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F519\r\n* docs: update .env.example with comprehensive variables and comments by @fancyboi999 in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F511\r\n* refactor: Consolidate backward compatibility into API models; simplify handler logic by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F520\r\n* Feat: add deepsearch agent for memos by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F517\r\n* add status of reasoning in playground by @Wang-Daoji in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F523\r\n* Feat: deepsearch agent dock search pipeline  by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F524\r\n* Feat\u002Fjoin test playground by @Wang-Daoji in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F527\r\n* Feat: update prompt and format need raw  by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F528\r\n* fix: cube init by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F529\r\n* Feat\u002Fredis scheduler by @glin93 in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F526\r\n* feat: add full text memory by jieba  by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F532\r\n* Feat： update code by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F533\r\n* fa_bu_hui pref by @Wang-Daoji in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F534\r\n* Feat\u002Fmerge api refactor to dev by @Wang-Daoji in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F531\r\n* fix: remove  by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F535\r\n* feat: enable multi-cube chat (read\u002Fwrite) & unify ChatRequest\u002FADDRequest normalization by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F521\r\n* feat: Multi-Model Memory Reader with Modular Parser Architecture by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F536\r\n* feat(qdrant):support qdrant cloud and add index by @fancyboi999 in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F522\r\n* Feat: remove dup func name and add agentic search by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F537\r\n* fix: Make from_","2025-12-24T13:19:44",{"id":258,"version":259,"summary_zh":260,"released_at":261},297672,"v1.1.3","## What's Changed\r\n* add: change deafult pre_load by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F338\r\n* feat: add memory size in product api by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F348\r\n* Feat: update load cubes by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F350\r\n* eat:reoganize prompt with reference in user content by @kakack in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F351\r\n* hotfix:noe4j community dataformat by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F353\r\n* milvus implement by @Wang-Daoji in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F354\r\n* fix: code ruff format by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F355\r\n* Fix\u002Fremove bug by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F356\r\n* Fix\u002Fapi client by @CarltonXiang in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F357\r\n* chore: bump version to v1.1.2 by @tangg555 in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F360\r\n* fix: remove old mem by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F361\r\n* feat: only single-db mode in nebula now; modify index gql for better efficiency by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F363\r\n* feat: add server api prd by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F362\r\n* Feat: add neo4j db for user_name by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F365\r\n* Feat: add chat complete for new server_api  by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F366\r\n* fix: nebula efficiency by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F369\r\n* Feat\u002Fmerge dev by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F374\r\n* Feat: sync test to dev by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F375\r\n* Eval scripts by @Nyakult in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F377\r\n* feat: merge dev into main and change api-server by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F376\r\n* fix: sqlite list users error by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F384\r\n* feat: introduce async memory add for TreeTextMemory using MemScheduler by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F373\r\n* add pm and pref eval scripts by @Nyakult in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F385\r\n* Meger update about scheduler and new api to Dev by @tangg555 in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F386\r\n* Feat\u002Fmerge inst cplt to dev by @Wang-Daoji in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F388\r\n* Feat: add reranker strategies and update configs by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F390\r\n* modify code in evaluation by @Wang-Daoji in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F392\r\n* fix bug in pref_mem return by @Wang-Daoji in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F399\r\n* add polardb by @wustzdy in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F395\r\n* feat: fix mode by @lijicode in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F400\r\n* Feat: remove long waring for internet and add content for memreader by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F401\r\n* feat: redis for sync history memories and new api of mixture search by @tangg555 in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F398\r\n* memos online api eval scripts and readme by @Nyakult in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F403\r\n* feat: fix sources by @lijicode in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F404\r\n* fix porlar by @lijicode in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F406\r\n* Feat\u002Farms by @CarltonXiang in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F402\r\n* Hotfix: memos playground prompt reverse by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F408\r\n* Feat\u002Fpref optimize update by @Wang-Daoji in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F409\r\n* feat: fix polardb graph by @lijicode in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F411\r\n* feat: async add api by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F410\r\n* use nacos by @lijicode in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F407\r\n* feat: async add api by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F413\r\n* revision of mixture api: add conversation turn and reduce 2 stage ranking to 1 stage by @tangg555 in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F405\r\n* Feat: add recall strategy  by @whipser030 in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F414\r\n* Revert \"Feat: add recall strategy \" by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F415\r\n* Feat： add new recall and verify  by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F416\r\n* Feat: remove usage data by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F417\r\n* feat: add moniter schedule by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F419\r\n* feat:turn off graph call by @whipser030 in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F418\r\n* pm & prefEval scripts updates by @Nyakult in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F421\r\n* add polardb pool by @wustzdy in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F420\r\n* Feat\u002Fpref optimize update by @Wang-Daoji in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F422\r\n* fix:tree file change Searcher inputs by @whipser030 in https:\u002F\u002Fgithub.com\u002FMemTensor\u002F","2025-11-07T03:24:45",{"id":263,"version":264,"summary_zh":265,"released_at":266},297673,"v1.1.2","## What's Changed\r\n* fix: api client get_message models by @CarltonXiang in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F359\r\n\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fcompare\u002Fv1.1.1...v1.1.2","2025-10-11T09:24:03",{"id":268,"version":269,"summary_zh":270,"released_at":271},297674,"v1.1.1","## What's Changed\r\n* feat: delete custom_logger_handler by @CarltonXiang in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F289\r\n* fix: change env model name by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F292\r\n* fix:#286:https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fissues\u002F286 by @kakack in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F293\r\n* Feat:add self defined memcube id for reg user by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F295\r\n* Feat\u002Fadd opentelmetry by @CarltonXiang in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F298\r\n* feat: add orginal context for reranking by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F284\r\n* revert: nebular require_python by @CarltonXiang in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F300\r\n* feat: chat bot api by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F294\r\n* feat: chat bot api by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F302\r\n* feat: chat bot api, add reranker filter; fix pydantic bug by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F303\r\n* fix: bug in internet pydantic error by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F304\r\n* Feat\u002Fadd opentelmetry by @CarltonXiang in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F307\r\n* feat: update nebula to nebula 5.1.1 by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F311\r\n* fix: nebula multi db bug by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F313\r\n* Feat\u002Fmemos client by @CarltonXiang in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F312\r\n* Feat: add time log for threaddict and change openai packacge singleton by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F314\r\n* Feat\u002Fadd timerlog by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F317\r\n* Feat\u002Fadd opentelmetry by @CarltonXiang in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F315\r\n* feat: add api client by @CarltonXiang in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F316\r\n* Feat: add segment lock dict  by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F319\r\n* fix:fix dump parallel for dumps cubes by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F320\r\n* feat: add sinlgleton by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F321\r\n* feat: nebula&reorganize update by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F322\r\n* fix: nebula reset bug by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F323\r\n* feat: add default processing in mem-reader by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F325\r\n* feat:add time step by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F326\r\n* Feat:add time step by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F327\r\n* docker start by @pursues in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F324\r\n* Feat: remove json by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F328\r\n* feat: remove  by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F329\r\n* fix: not include embedding by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F330\r\n* Feat\u002Fadd time step by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F331\r\n* Feat\u002Fadd time step by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F332\r\n* Fix\u002Fdefault add by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F333\r\n* feat: recall and searcher use parallel by @lijicode in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F337\r\n* Feat\u002Fapi client by @CarltonXiang in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F334\r\n* feat: API 1.0 by @CarltonXiang in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F339\r\n* fix: format by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F341\r\n* chore: bump version to v1.1.0 by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F340\r\n\r\n## New Contributors\r\n* @pursues made their first contribution in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F324\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fcompare\u002Fv1.0.1...v1.1.1","2025-09-24T17:15:36",{"id":273,"version":274,"summary_zh":275,"released_at":276},297675,"v1.1.0","## What's Changed\r\n* feat: delete custom_logger_handler by @CarltonXiang in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F289\r\n* fix: change env model name by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F292\r\n* fix:#286:https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fissues\u002F286 by @kakack in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F293\r\n* Feat:add self defined memcube id for reg user by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F295\r\n* Feat\u002Fadd opentelmetry by @CarltonXiang in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F298\r\n* feat: add orginal context for reranking by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F284\r\n* revert: nebular require_python by @CarltonXiang in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F300\r\n* feat: chat bot api by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F294\r\n* feat: chat bot api by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F302\r\n* feat: chat bot api, add reranker filter; fix pydantic bug by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F303\r\n* fix: bug in internet pydantic error by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F304\r\n* Feat\u002Fadd opentelmetry by @CarltonXiang in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F307\r\n* feat: update nebula to nebula 5.1.1 by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F311\r\n* fix: nebula multi db bug by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F313\r\n* Feat\u002Fmemos client by @CarltonXiang in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F312\r\n* Feat: add time log for threaddict and change openai packacge singleton by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F314\r\n* Feat\u002Fadd timerlog by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F317\r\n* Feat\u002Fadd opentelmetry by @CarltonXiang in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F315\r\n* feat: add api client by @CarltonXiang in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F316\r\n* Feat: add segment lock dict  by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F319\r\n* fix:fix dump parallel for dumps cubes by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F320\r\n* feat: add sinlgleton by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F321\r\n* feat: nebula&reorganize update by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F322\r\n* fix: nebula reset bug by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F323\r\n* feat: add default processing in mem-reader by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F325\r\n* feat:add time step by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F326\r\n* Feat:add time step by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F327\r\n* docker start by @pursues in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F324\r\n* Feat: remove json by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F328\r\n* feat: remove  by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F329\r\n* fix: not include embedding by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F330\r\n* Feat\u002Fadd time step by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F331\r\n* Feat\u002Fadd time step by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F332\r\n* Fix\u002Fdefault add by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F333\r\n* feat: recall and searcher use parallel by @lijicode in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F337\r\n* Feat\u002Fapi client by @CarltonXiang in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F334\r\n* feat: API 1.0 by @CarltonXiang in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F339\r\n* fix: format by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F341\r\n* chore: bump version to v1.1.0 by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F340\r\n\r\n## New Contributors\r\n* @pursues made their first contribution in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F324\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fcompare\u002Fv1.0.1...v1.1.0","2025-09-24T15:23:07",{"id":278,"version":279,"summary_zh":280,"released_at":281},297676,"v1.0.1","## What's Changed\r\n* feat: use different template for different language input by @Nyakult in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F232\r\n* fix: time hullucination by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F234\r\n* fix: chat time bug by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F235\r\n* push locomo rag eval code by @CSLiuPeng in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F180\r\n* feat: add further questions for dialogue by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F236\r\n* Fix: fix list user bugs and multi-user-examples get_all args by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F237\r\n* fix: nebula bug by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F242\r\n* Feat: change reference position and reorganize code by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F240\r\n* feat: reject answer by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F243\r\n* feat: support retrieval from specified memos_cube by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F244\r\n* feat: modify product answer prompt by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F245\r\n* feat: modify reference format by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F246\r\n* feat: memos add moscube turnoff by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F247\r\n* Fix: fix memcube path bug for docker and change further question prompt by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F248\r\n* Feat: add chat complete api for no-stream and  rewrite chat func for moscore by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F253\r\n* fix: mem-reader bug by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F255\r\n* feat: modify nebula session pool by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F259\r\n* fix: general_text add user_id by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F260\r\n* feat: Asynchronous processing of logs, notifications and memory addit… by @lijicode in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F261\r\n* feat: mos add load sdk for user by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F263\r\n* feat: enhance NebulaGraph pool management & improve Searcher usage logging by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F265\r\n* feat: add custom logger by @CarltonXiang in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F217\r\n* Feat: update chatbot for postprocessing memory by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F267\r\n* Feat\u002Fadd traceid by @CarltonXiang in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F270\r\n* feat: modify mem-reader prompt by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F273\r\n* Feat\u002Fadd traceid by @CarltonXiang in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F274\r\n* Feat: fix stream output and add openai stream by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F276\r\n* feat: add reranker by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F277\r\n* fix: reranker config bug by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F278\r\n* feat: adjust similarity threshold by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F279\r\n* feat: set minimun returned memories back to 3 by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F280\r\n* Feat: change mem prompt by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F281\r\n* feat: internet search speed and reranker by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F282\r\n* feat: update filter mem by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F285\r\n* feat: updatebug by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F287\r\n* feat: modify self intro by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F288\r\n* Chore: Change version to v1.0.1 by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F290\r\n* chore: bump version to v1.0.1 by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F275\r\n\r\n## New Contributors\r\n* @CSLiuPeng made their first contribution in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F180\r\n* @lijicode made their first contribution in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F261\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fcompare\u002Fv1.0.0...v1.0.1","2025-09-10T08:31:49",{"id":283,"version":284,"summary_zh":285,"released_at":286},297677,"v1.0.0","## What's Changed\r\n* update readme by @tangg555 in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F194\r\n* fix: nebula multi-embedding & add BochaAI Search Retriever by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F195\r\n* feat: modify product config by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F199\r\n* fix: fix bug when calling _concat_caches in kv.py (from pr#177) by @spitzblattr in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F205\r\n* Feat\u002Freorg dev by @Nyakult in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F202\r\n* feat: enhance retriever, reorganizer & NebulaGraph handling by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F200\r\n* feat & refactor: add search function feature to test scheduler on a modified locome benchmark, and slightly change the logic of query consume and query monitors by @tangg555 in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F204\r\n* feature & fix bugs: fix bugs after removing initialize_working_memory_monitors; add dispatcher_monitor designed to monitor the thread pool in the dispatcher moduler, and meanwhile dispatcher is enhanced with thread issue handlers by @tangg555 in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F207\r\n* feat: add custom request log by @CarltonXiang in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F206\r\n* fix: a temporal workaround for the \"virtualenv 20.33.0 breaks Poetry in GitHub Actions\" by @Ki-Seki in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F213\r\n* chore: The Poetry virtualenv issue persists; another temporary workaround is ok by @Ki-Seki in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F215\r\n* fix: reference for some string by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F211\r\n* fix: self.mem_cubes RuntimeError: dictionary changed size during iteration by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F216\r\n* feat: add embedding-aware graph query interfaces and parallel retrieval by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F212\r\n* feat: delete-custom-logger by @CarltonXiang in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F220\r\n* Refactor: Conflict resolution, bug fixes mainly for the mem scheduler by @tangg555 in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F221\r\n* feat: fix illumination by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F222\r\n* fix: reorganize and schedular bug by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F223\r\n* MemOS v0.2.3 release by @kakack in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F225\r\n* fix: remove field name will cause error for org-memscheduler config  by @fridayL in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F226\r\n* MemOS v0.2.3 release by @kakack in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F227\r\n* feat: illumination by @CaralHsi in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F229\r\n* Chore: Change version to v1.0.0 by @kakack in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F230\r\n* chore: bump version to v1.0.0 by @kakack in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F228\r\n\r\n## New Contributors\r\n* @spitzblattr made their first contribution in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F205\r\n* @CarltonXiang made their first contribution in https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fpull\u002F206\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002FMemTensor\u002FMemOS\u002Fcompare\u002Fv0.2.2...v1.0.0\r\n","2025-08-07T13:32:25"]