https://channel.nostses.icuhttps://channel.nostses.icu/posts/1907https://channel.nostses.icu/posts/1907Sun, 28 Dec 2025 01:32:40 GMT<div class="tgme_widget_message_forwarded_from accent_color">Forwarded from <a class="tgme_widget_message_forwarded_from_name" href="https://t.me/HardcoreOpenAI/1020" target="_blank" rel="noopener"><span>硬核开源智库</span></a></div><a href="https://github.com/llama-farm/llamafarm" target="_blank" rel="noopener" title="LlamaFarm">LlamaFarm</a> 是一个开源框架，专注于构建基于检索增强（RAG）和智能代理的AI应用。它内置了默认方案（本地模型 Ollama、向量存储 Chroma），但架构完全可扩展，支持随时替换运行时、数据库和解析器，无需重写代码。| <a href="/search/result?q=%23%E6%A1%86%E6%9E%B6" title="#框架">#框架</a><br /><br />主要特点：<br />- 本地优先体验，一条命令行工具（lf）管理项目、数据集和对话<br />- 生产级架构，支持配置驱动、模式校验的项目管理<br />- 灵活定制的RAG流水线，YAML配置即可轻松调整<br />- 集成多种AI运行时，支持本地模型和云端API无缝切换<br />- 丰富的CLI命令，实现项目初始化、数据上传、处理及聊天<br />- 提供OpenAI格式兼容的REST API，方便集成到各种应用<br /><br />支持 macOS、Linux 和 Windows，安装便捷，助力开发者快速搭建强大AI系统。<br /><a href="/search/result?q=%23%E8%B5%84%E6%BA%90%E5%8F%82%E8%80%83" title="#资源参考">#资源参考</a> <a href="/search/result?q=%23%E5%B7%A5%E5%85%B7" title="#工具">#工具</a> <a href="/search/result?q=%23RAG" title="#RAG">#RAG</a> <a href="/search/result?q=%23LLM" title="#LLM">#LLM</a><a class="tgme_widget_message_link_preview" href="https://github.com/llama-farm/llamafarm?utm_source=tldrai" target="_blank" rel="noopener" title="Deploy any AI model, agent, database, RAG, and pipeline locally or remotely in minutes - llama-farm/llamafarm"> <div class="link_preview_site_name accent_color">GitHub</div> <img class="link_preview_image" alt="GitHub - llama-farm/llamafarm: Deploy any AI model, agent, database, RAG, and pipeline locally or remotely in minutes" src="/static/https://cdn4.telesco.pe/file/H0_MVmeUn2WaYRvj29TTT3iPCtDQl7ZP7-uTZksaPrNYaK7CLFYdlMM1LHaOQcDVSN8agSaq07Ac7JNUrisl0So_rtq2vRki9dnJLWAE3nqD0mJvBdWmE7zaF5YLoaIcAPITdQLr3u2FW6alkfDDyz4xS6j8R6HZYJqyvX1EPB1mRJom-875kkZD6x5axxgXuntgbur0fETUBhaI4370-F3T2_BouPor4OcUSLDkbAZzldofCOvD5TT2l8S7Uk1rRlpNgKvp4EKnDI-U8naKNLBk3zxXUvGCjppM_7r_tPzaIYB0NUe3CgLm5bFWaiWoq8yOlU86qOAwsarr4pIJ6g.jpg" width="1200" height="630" loading="eager" /> <div class="link_preview_title">GitHub - llama-farm/llamafarm: Deploy any AI model, agent, database, RAG, and pipeline locally or remotely in minutes</div> <div class="link_preview_description">Deploy any AI model, agent, database, RAG, and pipeline locally or remotely in minutes - llama-farm/llamafarm</div> </a>https://channel.nostses.icu/posts/1904https://channel.nostses.icu/posts/1904Sun, 28 Dec 2025 01:32:40 GMT<div class="tgme_widget_message_forwarded_from accent_color">Forwarded from <a class="tgme_widget_message_forwarded_from_name" href="https://t.me/HardcoreOpenAI/1012" target="_blank" rel="noopener"><span>硬核开源智库</span></a></div><a href="https://github.com/openai/chatkit-js" target="_blank" rel="noopener" title="OpenAI ChatKit">OpenAI ChatKit</a> 是一个开箱即用的聊天框架，帮你快速集成高质量、AI 驱动的对话体验。无论是客服、智能助理还是社交应用，只需简单配置，就能拥有流畅自然的聊天界面和强大功能。| <a href="/search/result?q=%23%E6%A1%86%E6%9E%B6" title="#框架">#框架</a><br /><br />主要亮点：<br /><br />- 完全可定制的UI，自然融入你的应用<br />- 支持响应流，打造交互式会话体验<br />- 内置工具和工作流集成，支持智能推理和多步骤操作<br />- 丰富的交互组件，支持文件和图片上传<br />- 线程和消息管理，轻松组织复杂对话<br />- 来源注释与实体标注，保证透明度和引用准确<br /><br />只需几步，添加 ChatKit 组件，连接客户端令牌，就能快速上线，省时又高效。<a class="tgme_widget_message_link_preview" href="https://github.com/openai/chatkit-js" target="_blank" rel="noopener" title="Contribute to openai/chatkit-js development by creating an account on GitHub."> <div class="link_preview_site_name accent_color">GitHub</div> <div class="link_preview_title">GitHub - openai/chatkit-js</div> <div class="link_preview_description">Contribute to openai/chatkit-js development by creating an account on GitHub.</div> </a>https://channel.nostses.icu/posts/1901https://channel.nostses.icu/posts/1901Sun, 28 Dec 2025 01:32:40 GMT<div class="tgme_widget_message_forwarded_from accent_color">Forwarded from <a class="tgme_widget_message_forwarded_from_name" href="https://t.me/HardcoreOpenAI/1007" target="_blank" rel="noopener"><span>硬核开源智库</span></a></div><a href="https://github.com/PrimisAI/nexus" target="_blank" rel="noopener" title="PrimisAI Nexus">PrimisAI Nexus</a> 是一款强大灵活的 Python <mark class="highlight">框架</mark>，专为多智能体系统和大型语言模型（LLM）驱动的复杂任务编排设计。它支持创建多层级主管和助理主管结构，自动管理对话历史和日志，支持结构化输出和YAML配置，轻松实现任务分工和多智能体协作。| <a href="/search/result?q=%23%E6%A1%86%E6%9E%B6" title="#框架">#框架</a><br /><br />核心优势：<br /><br />- 统一管理多智能体，支持主控与助理主管分层协作 <br />- 内置持久化历史记录，保证会话连续性和复现性 <br />- 灵活配置LLM参数，支持自定义模型调用 <br />- 支持通过MCP协议集成远程或本地工具服务器，实现智能体功能扩展 <br />- 可使用YAML文件定义复杂智能体架构，方便快速部署和修改 <br />- 丰富调试与日志功能，助力开发与问题排查 <br /><br />适合 AI 研究者、开发者以及需要复杂自动化任务管理的团队。<br /><a href="/search/result?q=%23%E8%B5%84%E6%BA%90%E5%8F%82%E8%80%83" title="#资源参考">#资源参考</a> <a href="/search/result?q=%23AI%E6%A1%86%E6%9E%B6" title="#AI框架">#AI框架</a> #<a href="https://github.com/PrimisAI/nexus" target="_blank" rel="noopener" title="PrimisAI Nexus">PrimisAI Nexus</a><a class="tgme_widget_message_link_preview" href="https://github.com/PrimisAI/nexus" target="_blank" rel="noopener" title="A powerful Python framework for orchestrating AI agents and managing complex LLM-driven tasks with ease. - PrimisAI/nexus"> <div class="link_preview_site_name accent_color">GitHub</div> <img class="link_preview_image" alt="GitHub - PrimisAI/nexus: A powerful Python framework for orchestrating AI agents and managing complex LLM-driven tasks with ease." src="/static/https://cdn4.telesco.pe/file/BXxnA9fM1QPFGYXdJd5LQM0j1w1NZeZpreEecQFUtqfKT0R4FUsT5NX44pdJNjNW9ZwFQWOLm9F68dmnXev_LaNU1zivTreS6A_sU9Y2Zvpjc7BEzwFNLEEZvP_pJtxyethZiFkihTde7r308odVT9s2H2RuDV5qZ2S3ubIx8VRYlHZyG1t8ps1cim0V-QxGElV4QifhVIePwM_jQqPzMRLqKgNiYcHq4uBeeVVJWt-eg3rk6aA3FVQEB7ctcrNWYQgprJs0J_SY8sJLkavWasdQg1161WihhcPaNA0BRGPTtHgZOWSCssnGbxU-ihw91m4phUKbD548TmUZdYgBOQ.jpg" width="1200" height="630" loading="eager" /> <div class="link_preview_title">GitHub - PrimisAI/nexus: A powerful Python framework for orchestrating AI agents and managing complex LLM-driven tasks with ease.</div> <div class="link_preview_description">A powerful Python framework for orchestrating AI agents and managing complex LLM-driven tasks with ease. - PrimisAI/nexus</div> </a>https://channel.nostses.icu/posts/1874https://channel.nostses.icu/posts/1874Sun, 21 Dec 2025 11:50:40 GMT<div class="tgme_widget_message_forwarded_from accent_color">Forwarded from <a class="tgme_widget_message_forwarded_from_name" href="https://t.me/HardcoreOpenAI/982" target="_blank" rel="noopener"><span>硬核开源智库</span></a></div>在线开发搭建 web 应用时，常常要花很多时间写重复的代码，或者在 FastAPI 里集成 HTML 渲染很麻烦。<br /><br /><a href="https://github.com/feldroy/air" target="_blank" rel="noopener" title="Air">Air</a> 是由《Two Scoops of Django》作者打造的全新 Python Web <mark class="highlight">框架</mark>，基于 FastAPI、Starlette 和 Pydantic，专注于让你用 Python 优雅快速地写出网页和接口。| <a href="/search/result?q=%23%E6%A1%86%E6%9E%B6" title="#框架">#框架</a><br /><br />不仅支持高效的 HTML 代码生成（Air Tags），还能无缝混合 Jinja 模板，完美支持 HTMX，内置表单验证，极大提升开发体验。<br /><br />主要功能：<br /><br />- 结合 FastAPI，前后端同框开发，接口和网页一站式管理；<br />- Air Tags：用 Python 类写 HTML，类型安全且易维护；<br />- 支持 Jinja 模板和 Air Tags 混合渲染，灵活自由；<br />- 内置 HTMX 支持，轻松实现页面局部刷新和交互；<br />- Pydantic 驱动的表单验证，保证数据安全准确；<br />- 轻量，易学，文档详细，适合快速上手和大型项目。<br /><br />只需 <code>pip install -U air</code> 即可安装，适合想用 Python 快速构建现代 Web 应用的开发者。让你写代码更专注，效率更高，体验更佳<a class="tgme_widget_message_link_preview" href="https://github.com/feldroy/air" target="_blank" rel="noopener" title="The first web framework designed for AI to write. Built on Python, FastAPI, Pydantic, and HTMX. By the authors of Two Scoops of Django. - feldroy/air"> <div class="link_preview_site_name accent_color">GitHub</div> <img class="link_preview_image" alt="GitHub - feldroy/air: The first web framework designed for AI to write. Built on Python, FastAPI, Pydantic, and HTMX. By the authors…" src="/static/https://cdn4.telesco.pe/file/F_as5VAKRMrxdvErNoPL-TMFU4YXDcJUaFKwANeLic-VC2-a-ZmioltMZmq6qSQUINhfdayTUaZtMzG4JVVZiXKW8mSSw09wlLhAJqpEZvkWi5tZoPVBbjKc-TYetGYLB7i-xIaLs15XfluKfba3X7h-2CMQlPPlooc_aLLL6oLi-JvosdrAMIwuk-BcKCKzI2GtkTd6etaWyKUEDceMh_CjMM2qrcWXYCelO-ck6bMYBs6cHixli3xolcLC4Wz48zx2tfJNIheu5hYlBzzFhzl4QkUXbfCL_u8Tz_CelbocUxVXCiyRD9mfSB57Cl1u0c--L1Q7XOakX-U0jpOR5Q.jpg" width="1200" height="630" loading="eager" /> <div class="link_preview_title">GitHub - feldroy/air: The first web framework designed for AI to write. Built on Python, FastAPI, Pydantic, and HTMX. By the authors…</div> <div class="link_preview_description">The first web framework designed for AI to write. Built on Python, FastAPI, Pydantic, and HTMX. By the authors of Two Scoops of Django. - feldroy/air</div> </a>https://channel.nostses.icu/posts/1433https://channel.nostses.icu/posts/1433Thu, 21 Aug 2025 19:28:23 GMT<div class="tgme_widget_message_forwarded_from accent_color">Forwarded from <a class="tgme_widget_message_forwarded_from_name" href="https://t.me/HardcoreOpenAI/817" target="_blank" rel="noopener"><span>硬核开源智库</span></a></div><a href="https://github.com/interval-package/trackerLab" target="_blank" rel="noopener" title="TrackerLab">TrackerLab</a>：集成IsaacLab与多模态全身控制的模块化框架，专为复杂人形机器人动作设计打造。 | <a href="/search/result?q=%23%E6%A1%86%E6%9E%B6" title="#框架">#框架</a><br /><br />• <i class="emoji"><b>🦿</b></i> 完全集成IsaacLab，采用管理器架构简化运动追踪与控制 <br />• <i class="emoji"><b>🔁</b></i> 支持SMPL/AMASS/FBX动作数据的全链路重定向，包含T姿态对齐、滤波与插值 <br />• <i class="emoji"><b>🎮</b></i> 多种控制模式自由切换，包括外骨骼姿态控制、PHC等，指令管理灵活高效 <br />• <i class="emoji"><b>🔀</b></i> 基于有限状态机（FSM）的技能图设计，支持手动触发、规划器和摇杆操作 <br />• <i class="emoji"><b>⚙️</b></i> 支持Unitree H1等真实机器人，代码清晰、易扩展，方便集成新动作与控制策略 <br />• <i class="emoji"><b>📂</b></i> 自带定制Gym环境，可直接用于IsaacLab训练脚本，快速上手多任务训练与播放 <br />• <i class="emoji"><b>📚</b></i> 提供中英文教程、完整数据集准备指南，零依赖独立安装，快速部署体验 <br />• <i class="emoji"><b>💼</b></i> 开源MIT许可，欢迎研究者与开发者合作共建，推动人形机器人智能运动前沿<a class="tgme_widget_message_link_preview" href="https://github.com/interval-package/trackerLab" target="_blank" rel="noopener" title="TrackerLab is a cutting-edge modular framework for humanoid motion retargeting, trajectory tracking, and skill-level control, built on top of IsaacLab. - Renforce-Dynamics/trackerLab"> <div class="link_preview_site_name accent_color">GitHub</div> <div class="link_preview_title">GitHub - Renforce-Dynamics/trackerLab: TrackerLab is a cutting-edge modular framework for humanoid motion retargeting, trajectory…</div> <div class="link_preview_description">TrackerLab is a cutting-edge modular framework for humanoid motion retargeting, trajectory tracking, and skill-level control, built on top of IsaacLab. - Renforce-Dynamics/trackerLab</div> </a>https://channel.nostses.icu/posts/1227https://channel.nostses.icu/posts/1227Wed, 30 Jul 2025 17:46:08 GMT<div class="tgme_widget_message_forwarded_from accent_color">Forwarded from <a class="tgme_widget_message_forwarded_from_name" href="https://t.me/HardcoreOpenAI/704" target="_blank" rel="noopener"><span>硬核开源智库</span></a></div>多Agent协同提升代码质量的自动化框架<br /><br />• 并行运行20+ Claude Code agents，最高支持50个，极大加速大规模代码库的自动化改进 <br />• 三大工作流支持：传统Bug修复、系统化最佳实践落地、多Agent协同复杂开发 <br />• 先进锁机制保障多Agent无冲突作业，自动管理任务分配和状态同步，避免重复和冲突 <br />• 涵盖34种主流技术栈（Next.js、Python、Rust、Go、Java、Flutter等），配置丰富，适配广泛项目 <br />• 实时监控面板直观显示Agent状态、上下文使用率、心跳频率及错误信息，支持tmux多视图操作 <br />• 自动恢复与上下文管理：Agent异常自动重启，自动清理上下文防止状态溢出，支持一键广播清理命令 <br />• 配套24套环境搭建脚本，覆盖从前端、后端到DevOps及数据工程，极大简化开发环境配置 <br />• 细粒度Git集成：支持定制分支、增量提交与详尽HTML运行报告，保证代码变更透明且易追踪 <br />• 灵活JSON配置系统，支持变量替换与动态块大小调整，满足不同项目和团队的个性化需求 <br />• 安全稳健：自动备份与恢复Claude设置，文件锁定防止并发冲突，异常清理确保环境干净 <br /><br />该框架不仅是多Agent任务调度工具，更是面向大规模代码库的智能协同开发平台，充分利用AI并行能力与严格的作业协调机制，解决传统自动化工具难以兼顾的代码冲突与任务重复问题，极大提高代码质量和团队开发效率，具备长期演进和多场景复用价值。<br /><br /><a href="https://github.com/Dicklesworthstone/claude_code_agent_farm" target="_blank" rel="noopener" title="Claude Code Agent Farm">Claude Code Agent Farm</a> | <a href="/search/result?q=%23%E6%A1%86%E6%9E%B6" title="#框架">#框架</a> <a href="/search/result?q=%23%E8%B5%84%E6%BA%90%E5%8F%82%E8%80%83" title="#资源参考">#资源参考</a> <a href="/search/result?q=%23Agent" title="#Agent">#Agent</a><a class="tgme_widget_message_link_preview" href="https://github.com/Dicklesworthstone/claude_code_agent_farm" target="_blank" rel="noopener" title="Orchestration framework for running 20+ Claude Code agents in parallel: automated bug fixing, best-practices sweeps, lock-based coordination, and real-time tmux monitoring - Dicklesworthstone/claud..."> <div class="link_preview_site_name accent_color">GitHub</div> <img class="link_preview_image" alt="GitHub - Dicklesworthstone/claude_code_agent_farm: Orchestration framework for running 20+ Claude Code agents in parallel: automated…" src="/static/https://cdn4.telesco.pe/file/khKigMYeNbyI8lWABIEWQDoWdv0bW8OYic0VXf3yT9JE1XYcOjjfRSxyTP86tK0-PxlQKSp7hf-zoshmtHzVCIwTRlqABFeNP1_D8WMtON7zcP50xVbwPlLAcVHy96rQ4CgeOxsP7vCfX9jx6ypA9CN9zq5bte4zbin-8sNTgWhWvu2o3Wx93CzCoB6_jXIsA08EesnsldBMBQF8OdFJ9PphXVwOS1rg7kxTzN5iK63OqcBpF58lUJSwhSIacVY_a4E6n1l_WN9GzAdjeD60JLIgb69u40wf7F6XumEh1JK1N1IuxCuNEbqNczJ-DWCR6mXjpOJKZDlmkViiqyVxcw.jpg" width="1200" height="630" loading="eager" /> <div class="link_preview_title">GitHub - Dicklesworthstone/claude_code_agent_farm: Orchestration framework for running 20+ Claude Code agents in parallel: automated…</div> <div class="link_preview_description">Orchestration framework for running 20+ Claude Code agents in parallel: automated bug fixing, best-practices sweeps, lock-based coordination, and real-time tmux monitoring - Dicklesworthstone/claud...</div> </a>https://channel.nostses.icu/posts/1204https://channel.nostses.icu/posts/1204Wed, 30 Jul 2025 17:35:38 GMT<div class="tgme_widget_message_forwarded_from accent_color">Forwarded from <a class="tgme_widget_message_forwarded_from_name" href="https://t.me/HardcoreOpenAI/635" target="_blank" rel="noopener"><span>硬核开源智库</span></a></div>面向开发者和内容创作者的先进视频制作框架，专注于用代码驱动视频生成与编辑，提升工作流效率与创意自由度。<i class="emoji"><b>🎥</b></i><br /><br />• 基于 React，支持用熟悉的前端技术构建高质量视频组件，实现动态内容与交互式设计。 <br />• 提供完整渲染管线，兼容浏览器和服务器端渲染，确保视频输出高效且稳定。 <br />• 支持多种格式导出，灵活适配不同平台需求，覆盖社交媒体、广告及教育视频制作场景。 <br />• 内置时间轴和动画控制，便于精细调整视频节奏和视觉效果，提升内容表现力。 <br />• 强调代码驱动的创作流程，促进版本管理、团队协作及自动化生产，适合规模化视频项目。 <br />• 开源生态活跃，拥有丰富插件与社区支持，持续优化功能与性能。 <br /><br />Remotion通过代码与视频的结合，打破传统编辑限制，推动视频制作进入高度模块化与自动化时代。适合追求效率与创新的专业团队长期参考与应用。<br /><br /><a href="https://www.remotion.dev/" target="_blank" rel="noopener" title="Remotion">Remotion</a> | <a href="/search/result?q=%23%E6%A1%86%E6%9E%B6" title="#框架">#框架</a><a class="tgme_widget_message_link_preview" href="https://www.remotion.dev/" target="_blank" rel="noopener"> <div class="link_preview_site_name accent_color">www.remotion.dev</div> <img class="link_preview_image" alt="Remotion | Make videos programmatically" src="/static/https://cdn4.telesco.pe/file/GkKgJVlnzVCgcyIU1mV2dSN1nSQevcIr0tzQWxnMRuqg11JJr3TPYE4_jTZS-6fuxTedDrj4gIgrWZRKV97zZ7oDf_oG50GG0sjyJT3WwkqWmWKOAI504EGjRDbB2-SUgnp5AQUBEUw21Ntszct9_HGB75PIiIxUAuHsI2kfysIHW_OnbxhNvB1oLZEcHfOZK9ggXTjNszirxf0poqCYQM3RtkzQQmCR1iKSRkZ5Slf2uQqaaUw2qi1odTyYSzMT9zXhZ0yEHqleXzWv3DiQFfn61Jlcd4buEIpHGcNVNdAYzz3KFxyPfCUPqSGq9XTGUicEfmQHWnkZ2WAXHrFxCQ.jpg" width="1200" height="630" loading="eager" /> <div class="link_preview_title">Remotion | Make videos programmatically</div> </a>https://channel.nostses.icu/posts/1202https://channel.nostses.icu/posts/1202Wed, 30 Jul 2025 17:35:38 GMT<div class="tgme_widget_message_forwarded_from accent_color">Forwarded from <a class="tgme_widget_message_forwarded_from_name" href="https://t.me/HardcoreOpenAI/633" target="_blank" rel="noopener"><span>硬核开源智库</span></a></div>专为开发者优化的 Claude Code 扩展框架，融合专用命令、智能角色与 MCP 服务器，助力高效开发流程。<br /><br />• 16 条高频开发命令覆盖实现、构建、设计、分析、测试、文档等核心环节，提升任务执行效率<i class="emoji"><b>🛠</b></i> <br />• 多领域智能角色（架构师、前端、后端、安全等）自动匹配专家视角，精准应对复杂场景<i class="emoji"><b>🎭</b></i> <br />• MCP 服务器集成支持官方文档调用、UI 组件生成、浏览器自动化，扩展能力显著提升<i class="emoji"><b>🔧</b></i> <br />• 统一 CLI 安装器和多种安装方式，兼容 Python 3.8+，支持跨平台 uv / uvx 快速部署 <br />• 任务管理与 Token 优化机制，保障长会话下的上下文连贯与资源节省 <br />• v3 架构更简洁、性能更优，移除钩子系统以待 v4 重新设计，持续迭代中 <br /><br />SuperClaude 通过模块化设计与智能路由，将 AI 助手从泛用工具转变为开发者的多面能手，提升协同效率与专业深度。 <br /><br />长期价值：架构调整体现对稳定性与可扩展性的深刻理解，代码与文档开放，适合持续贡献与社区共建。 <br /><br /><a href="https://github.com/SuperClaude-Org/SuperClaude_Framework" target="_blank" rel="noopener" title="SuperClaude v3">SuperClaude v3</a> | <a href="/search/result?q=%23%E6%A1%86%E6%9E%B6" title="#框架">#框架</a><a class="tgme_widget_message_link_preview" href="https://github.com/SuperClaude-Org/SuperClaude_Framework" target="_blank" rel="noopener" title="A configuration framework that enhances Claude Code with specialized commands, cognitive personas, and development methodologies. - SuperClaude-Org/SuperClaude_Framework"> <div class="link_preview_site_name accent_color">GitHub</div> <img class="link_preview_image" alt="GitHub - SuperClaude-Org/SuperClaude_Framework: A configuration framework that enhances Claude Code with specialized commands,…" src="/static/https://cdn4.telesco.pe/file/WrYD1AwNnIm-cKuy2cuXjszKMx4kFKgeUVLk5jJL_FhwiYf9pOk4g3RKMhaUXkUgaVavtkfL6SYwg20XJB6Z42crXx-B_dKl2l9_KDd0sQyvJY7g--d7T39b1DPXYkxS2mYq5lpFTiDTUHFm375dLTOuZu2paz0lHfvltejOALmjEeWvop2zV3elB40bztpFp1tRH70b8kTgyJBrxNOS6A091-UOld8IaVQGKKQyOV0ilhgSGKb2ksb_kMG9RTf7lf3Qs68Qk7c0BEfQa859TfhTBptB56EhDB7oUntzHZbIWsl7Y--JpydrpVEHPkKTHRs5hk-Q6OfbFXYzKP5KqA.jpg" width="1200" height="630" loading="eager" /> <div class="link_preview_title">GitHub - SuperClaude-Org/SuperClaude_Framework: A configuration framework that enhances Claude Code with specialized commands,…</div> <div class="link_preview_description">A configuration framework that enhances Claude Code with specialized commands, cognitive personas, and development methodologies. - SuperClaude-Org/SuperClaude_Framework</div> </a>https://channel.nostses.icu/posts/1172https://channel.nostses.icu/posts/1172Wed, 30 Jul 2025 05:29:26 GMT<div class="tgme_widget_message_forwarded_from accent_color">Forwarded from <a class="tgme_widget_message_forwarded_from_name" href="https://t.me/HardcoreOpenAI/563" target="_blank" rel="noopener"><span>硬核开源智库</span></a></div><a href="https://github.com/getzep/graphiti" target="_blank" rel="noopener" title="Graphiti">Graphiti</a>：实时构建面向AI智能体的知识图谱框架，专为动态环境设计，支持连续增量更新、双时间维度管理及高效混合检索，极大提升AI记忆与推理能力。| <a href="/search/result?q=%23%E6%A1%86%E6%9E%B6" title="#框架">#框架</a> <a href="/search/result?q=%23%E6%99%BA%E8%83%BD%E4%BD%93" title="#智能体">#智能体</a> <a href="/search/result?q=%23Agent" title="#Agent">#Agent</a> <a href="/search/result?q=%23%E8%B5%84%E6%BA%90%E5%8F%82%E8%80%83" title="#资源参考">#资源参考</a> <br /><br />核心亮点：<br />• 实时整合用户交互、结构化与非结构化数据，动态维护知识图谱<br />• 支持语义、关键词（BM25）与图遍历混合检索，响应延迟低至亚秒级<br />• 自定义实体定义，灵活构建适合业务的本体模型<br />• 双时间模型，精准记录事件发生与数据摄取时间，支持历史时点查询<br />• 高度可扩展，支持Neo4j、FalkorDB等后端，适合企业级大规模数据环境<br />• 开源免费，基于Apache 2.0协议，社区活跃，持续迭代优化<br /><br />应用场景：<br />• AI助手记忆管理与上下文推理<br />• 动态企业数据整合与知识管理<br />• 多模态复杂查询与任务自动化<br /><br />快速上手：<br />Python 3.10+，搭配Neo4j或FalkorDB，默认集成OpenAI LLM及嵌入服务，支持Google Gemini、Anthropic、Groq及本地Ollama模型。<a class="tgme_widget_message_link_preview" href="https://github.com/getzep/graphiti" target="_blank" rel="noopener" title="Build Real-Time Knowledge Graphs for AI Agents. Contribute to getzep/graphiti development by creating an account on GitHub."> <div class="link_preview_site_name accent_color">GitHub</div> <img class="link_preview_image" alt="GitHub - getzep/graphiti: Build Real-Time Knowledge Graphs for AI Agents" src="/static/https://cdn4.telesco.pe/file/pc_742rgt8YQdyIw0MRN94ObQ9tlKkwrDINiOaXY_DQqCMT7MJUHclIV4iNTpTULhHH8dEwjTfZ1DfPR0sBkuhx-awOHHCCgw7YXBCOFchzQ7879WTa_AJ0gGrFewI8DFYxjZ7iOIUF_cbN4sQwDMKu5zQznFvEDJetRHZcxz6HZp39gFR_otusQhT7sLFghrJx4mPH2kOkmzxM4QbvqRHmCxu-YXB_8peVJ-sp4wzdDXTkZI2deTtKB_G3whHmxLSYugFKl2IFfA7SaDJF86leaWsZAx0BgEFRA8CO7A6PGEO69VxEirLaqNNYJyFWIJv6iQBhqA-T307-LO40xbw.jpg" width="1200" height="630" loading="eager" /> <div class="link_preview_title">GitHub - getzep/graphiti: Build Real-Time Knowledge Graphs for AI Agents</div> <div class="link_preview_description">Build Real-Time Knowledge Graphs for AI Agents. Contribute to getzep/graphiti development by creating an account on GitHub.</div> </a>https://channel.nostses.icu/posts/1073https://channel.nostses.icu/posts/1073Thu, 10 Jul 2025 05:27:35 GMT<div class="tgme_widget_message_forwarded_from accent_color">Forwarded from <a class="tgme_widget_message_forwarded_from_name" href="https://t.me/HardcoreOpenAI/503" target="_blank" rel="noopener"><span>硬核开源智库</span></a></div>为AI和Agent应用量身打造的DataFrame框架，让数据处理更智能、更高效。<br /><br />专为LLM推理设计，支持自动批量优化和API调用重试；提供丰富的语义操作符，如情感分析、文本分类、数据关联；原生支持Markdown、转录文本和复杂JSON处理<br /><br /><a href="https://github.com/typedef-ai/fenic" target="_blank" rel="noopener" title="fenic">fenic</a> | <a href="/search/result?q=%23%E6%A1%86%E6%9E%B6" title="#框架">#框架</a><a class="tgme_widget_message_link_preview" href="https://github.com/typedef-ai/fenic?utm_source=tldrnewsletter" target="_blank" rel="noopener" title="Semantic DataFrames for humans and agents. Contribute to typedef-ai/fenic development by creating an account on GitHub."> <div class="link_preview_site_name accent_color">GitHub</div> <img class="link_preview_image" alt="GitHub - typedef-ai/fenic: Semantic DataFrames for humans and agents" src="/static/https://cdn4.telesco.pe/file/qf_VaLMbWPX4x-fTZ3TktlXen28nM4nAwlWkZ6iElSPM5whHgTQsgsefXhOnzlq_PXefVWJHEBFYLD32XLlo0C3I6dZBW0IEaW5tR5exBl2GdHvgUiCxPjQWAXs86LfvxawJBSGoXw0q4DvLixNskBGZF7tCQtgcnkb4-8nn86LAJLj7YY0oy69Y151_jn8AKT38Vl2biQ5I8SeGYTOTUNjx5BoN7jK10o3md36XqYXQBNoQei9ikin1Wwy_p-4HTXD6a74lNCOawRxSNc3R4oo9SO2SsEwBqchPOvlbqW6BpNBKeeHVPEppAHuveirv6hKbjDmXLtLl5JoB2to6cw.jpg" width="1200" height="630" loading="eager" /> <div class="link_preview_title">GitHub - typedef-ai/fenic: Semantic DataFrames for humans and agents</div> <div class="link_preview_description">Semantic DataFrames for humans and agents. Contribute to typedef-ai/fenic development by creating an account on GitHub.</div> </a>