[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"similar-NeumTry--NeumAI":3,"tool-NeumTry--NeumAI":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",159636,2,"2026-04-17T23:33:34",[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":78,"owner_twitter":79,"owner_website":80,"owner_url":81,"languages":82,"stars":87,"forks":88,"last_commit_at":89,"license":90,"difficulty_score":32,"env_os":91,"env_gpu":91,"env_ram":91,"env_deps":92,"category_tags":97,"github_topics":99,"view_count":32,"oss_zip_url":77,"oss_zip_packed_at":77,"status":17,"created_at":117,"updated_at":118,"faqs":119,"releases":149},8930,"NeumTry\u002FNeumAI","NeumAI","Neum AI is a best-in-class framework to manage the creation and synchronization of vector embeddings at large scale.","Neum AI 是一款专为大规模向量嵌入（Vector Embeddings）管理打造的开源框架，旨在帮助开发者高效构建基于检索增强生成（RAG）的大语言模型应用。它打通了从数据源提取、内容处理到向量数据库入库的全流程，让非结构化数据能轻松转化为大模型可理解的上下文信息。\n\n面对海量数据处理时，传统方案往往在数据连接器集成、嵌入模型调用及向量存储同步上耗费大量精力。Neum AI 通过提供内置的多种数据源连接器、嵌入服务接口及向量存储适配，显著降低了集成复杂度。其核心优势在于高吞吐量的分布式架构，能够并行处理数十亿级数据点，并支持数据源的实时同步，确保知识库始终最新。此外，它还提供了灵活的数据预处理能力（如加载、分块、筛选）以及自动化的元数据追踪，助力实现更精准的混合检索。\n\n这款工具特别适合需要构建企业级 RAG 应用的软件开发者和数据工程师。无论是希望快速验证原型的初创团队，还是需处理百万级文档的大型项目，Neum AI 都能提供从本地开发到云端部署的一站式解决方案，让用户将更多时间专注于业务逻辑而非底层基础设施的搭建。","\u003Ch1 align=\"center\">Neum AI\u003C\u002Fh1>\n\n\u003Cdiv align=\"center\">\n  \n  [Homepage](https:\u002F\u002Fwww.neum.ai) | [Documentation](https:\u002F\u002Fdocs.neum.ai) | [Blog](https:\u002F\u002Fneum.ai\u002Fblog) | [Discord](https:\u002F\u002Fdiscord.gg\u002FmJeNZYRz4m) | [Twitter](https:\u002F\u002Ftwitter.com\u002Fneum_ai)\n  \n  \u003Ca href=\"https:\u002F\u002Fwww.ycombinator.com\u002Fcompanies\u002Fneum-ai\">\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FNeumTry_NeumAI_readme_64e8b3dee6f7.png\"\u002F>\u003C\u002Fa> \n  \u003Ca href=\"https:\u002F\u002Fpypi.org\u002Fproject\u002Fneumai\u002F\">\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fpypi\u002Fv\u002Fneumai\" alt=\"PyPI\">\n  \u003C\u002Fa>\n\u003C\u002Fdiv>\n\n![Neum AI Hero](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FNeumTry_NeumAI_readme_0ec7eeb1b4a4.png)\n\n**[Neum AI](https:\u002F\u002Fneum.ai) is a data platform that helps developers leverage their data to contextualize Large Language Models through Retrieval Augmented Generation (RAG)** This includes\nextracting data from existing data sources like document storage and NoSQL, processing the contents into vector embeddings and ingesting the vector embeddings into vector databases for similarity search. \n\nIt provides you a comprehensive solution for RAG that can scale with your application and reduce the time spent integrating services like data connectors, embedding models and vector databases.\n\n## Features\n\n- 🏭 **High throughput distributed architecture** to handle billions of data points. Allows high degrees of parallelization to optimize embedding generation and ingestion.\n- 🧱 **Built-in data connectors** to common data sources, embedding services and vector stores.\n- 🔄 **Real-time synchronization** of data sources to ensure your data is always up-to-date. \n- ♻ **Customizable data pre-processing** in the form of loading, chunking and selecting.\n- 🤝 **Cohesive data management** to support hybrid retrieval with metadata. Neum AI automatically augments and tracks metadata to provide rich retrieval experience.\n\n## Talk to us\n\nYou can reach our team either through email ([founders@tryneum.com](mailto:founders@tryneum.com)), on [discord](https:\u002F\u002Fdiscord.gg\u002FmJeNZYRz4m) or by [scheduling a call wit us](https:\u002F\u002Fcalendly.com\u002Fneum-ai\u002Fneum-ai-demo?month=2023-12).\n\n## Getting Started\n\n### Neum AI Cloud\n\nSign up today at [dashboard.neum.ai](https:\u002F\u002Fdashboard.neum.ai). See our [quickstart](https:\u002F\u002Fdocs.neum.ai\u002Fget-started\u002Fquickstart) to get started.\n\nThe Neum AI Cloud supports a large-scale, distributed architecture to run millions of documents through vector embedding. For the full set of features see: [Cloud vs Local](https:\u002F\u002Fneumai.mintlify.app\u002Fget-started\u002Fcloud-vs-local)\n\n### Local Development\n\nInstall the [`neumai`](https:\u002F\u002Fpypi.org\u002Fproject\u002Fneumai\u002F) package:\n\n```bash\npip install neumai\n```\n\nTo create your first data pipelines visit our [quickstart](https:\u002F\u002Fdocs.neum.ai\u002Fget-started\u002Fquickstart).\n\nAt a high level, a pipeline consists of one or multiple sources to pull data from, one embed connector to vectorize the content, and one sink connector to store said vectors.\nWith this snippet of code we will craft all of these and run a pipeline:\n\u003Cdetails open>\u003Csummary>\n  \n  ### Creating and running a pipeline\n  \u003C\u002Fsummary>\n  \n  ```python\n  \n  from neumai.DataConnectors.WebsiteConnector import WebsiteConnector\n  from neumai.Shared.Selector import Selector\n  from neumai.Loaders.HTMLLoader import HTMLLoader\n  from neumai.Chunkers.RecursiveChunker import RecursiveChunker\n  from neumai.Sources.SourceConnector import SourceConnector\n  from neumai.EmbedConnectors import OpenAIEmbed\n  from neumai.SinkConnectors import WeaviateSink\n  from neumai.Pipelines import Pipeline\n\n  website_connector =  WebsiteConnector(\n      url = \"https:\u002F\u002Fwww.neum.ai\u002Fpost\u002Fretrieval-augmented-generation-at-scale\",\n      selector = Selector(\n          to_metadata=['url']\n      )\n  )\n  source = SourceConnector(\n      data_connector = website_connector, \n      loader = HTMLLoader(), \n      chunker = RecursiveChunker()\n  )\n\n  openai_embed = OpenAIEmbed(\n      api_key = \"\u003COPEN AI KEY>\",\n  )\n\n  weaviate_sink = WeaviateSink(\n      url = \"your-weaviate-url\",\n      api_key = \"your-api-key\",\n      class_name = \"your-class-name\",\n  )\n\n  pipeline = Pipeline(\n      sources=[source], \n      embed=openai_embed, \n      sink=weaviate_sink\n  )\n  pipeline.run()\n\n  results = pipeline.search(\n      query=\"What are the challenges with scaling RAG?\", \n      number_of_results=3\n  )\n\n  for result in results:\n      print(result.metadata)\n  ```\n\u003C\u002Fdetails>\n\n\u003Cdetails>\u003Csummary>\n\n  ### Creating and running a pipeline - Postgres connector\n  \u003C\u002Fsummary>\n\n  ```python\n  \n  from neumai.DataConnectors.PostgresConnector import PostgresConnector\n  from neumai.Shared.Selector import Selector\n  from neumai.Loaders.JSONLoader import JSONLoader\n  from neumai.Chunkers.RecursiveChunker import RecursiveChunker\n  from neumai.Sources.SourceConnector import SourceConnector\n  from neumai.EmbedConnectors import OpenAIEmbed\n  from neumai.SinkConnectors import WeaviateSink\n  from neumai.Pipelines import Pipeline\n\n  website_connector =  PostgresConnector(\n      connection_string = 'postgres',\n      query = 'Select * from ...'\n  )\n  source = SourceConnector(\n      data_connector = website_connector, \n      loader = JSONLoader(\n          id_key='\u003Cyour id key of your jsons>',\n          selector=Selector(\n              to_embed=['property1_to_embed','property2_to_embed'],\n              to_metadata=['property3_to_include_in_metadata_in_vector']\n          )\n      ),\n      chunker = RecursiveChunker()\n  )\n\n  openai_embed = OpenAIEmbed(\n      api_key = \"\u003COPEN AI KEY>\",\n  )\n\n  weaviate_sink = WeaviateSink(\n      url = \"your-weaviate-url\",\n      api_key = \"your-api-key\",\n      class_name = \"your-class-name\",\n  )\n\n  pipeline = Pipeline(\n      sources=[source], \n      embed=openai_embed, \n      sink=weaviate_sink\n  )\n\n  pipeline.run()\n\n  results = pipeline.search(\n      query=\"...\", \n      number_of_results=3\n  )\n\n  for result in results:\n      print(result.metadata)\n  ```\n\u003C\u002Fdetails>\n\n\u003Cdetails>\u003Csummary>\n  \n  ### Publishing pipeline to Neum Cloud\n  \u003C\u002Fsummary>\n  \n  ```python\n  from neumai.Client.NeumClient import NeumClient\n  client = NeumClient(\n      api_key='\u003Cyour neum api key, get it from https:\u002F\u002Fdashboard.neum.ai',\n  )\n  client.create_pipeline(pipeline=pipeline)\n  ```\n\u003C\u002Fdetails>\n\n### Self-host\n\nIf you are interested in deploying Neum AI to your own cloud contact us at [founders@tryneum.com](mailto:founders@tryneum.com).\n\nWe have a sample backend architecture published on [GitHub](https:\u002F\u002Fgithub.com\u002FNeumTry\u002Fneum-at-scale) which you can use as a starting point.\n\n## Available Connectors\nFor an up-to-date list please visit our [docs](https:\u002F\u002Fdocs.neum.ai\u002Fcomponents\u002FsourceConnector)\n\n\u003Cdetails>\n\n### Source connectors\n1. Postgres\n2. Hosted Files\n3. Websites\n4. S3\n5. Azure Blob\n6. Sharepoint\n7. Singlestore\n8. Supabase Storage\n\n### Embed Connectors\n1. OpenAI embeddings\n2. Azure OpenAI embeddings\n\n### Sink Connectors\n1. Supabase postgres\n2. Weaviate\n3. Qdrant\n4. Pinecone\n5. Singlestore\n\n\u003C\u002Fdetails>\n\n## Roadmap\nOur roadmap is evolving with asks, so if there is anything missing feel free to open an issue or send us a message.\n\n\u003Cdetails>\n  \nConnectors\n- [ ]  MySQL - Source\n- [ ]  GitHub - Source\n- [ ]  Google Drive - Source\n- [ ]  Hugging Face - Embedding\n- [x]  LanceDB - Sink\n- [x]  Marqo - Sink\n- [ ]  Milvus - Sink\n- [ ]  Chroma - Sink\n\nSearch\n- [x]  Retrieval feedback\n- [x]  Filter support\n- [x]  Unified Neum AI filters\n- [ ]  Smart routing (w\u002F embedding based classification)\n- [ ]  Smart routing (w\u002F LLM based classification)\n- [ ]  Self-Query Retrieval (w\u002F Metadata attributes generation)\n\nExtensibility\n- [x]  Langchain \u002F Llama Index Document to Neum Document converter\n- [ ]  Custom chunking and loading\n\nExperimental\n- [ ]  Async metadata augmentation\n- [ ]  Chat history connector\n- [ ]  Structured (SQL and GraphQL) search connector\n\u003C\u002Fdetails>\n\n## Neum Tools\nAdditional tooling for Neum AI can be found here:\n\n- [neumai-tools](https:\u002F\u002Fpypi.org\u002Fproject\u002Fneumai-tools\u002F): contains pre-processing tools for loading and chunking data before generating vector embeddings.\n","\u003Ch1 align=\"center\">Neum AI\u003C\u002Fh1>\n\n\u003Cdiv align=\"center\">\n  \n  [首页](https:\u002F\u002Fwww.neum.ai) | [文档](https:\u002F\u002Fdocs.neum.ai) | [博客](https:\u002F\u002Fneum.ai\u002Fblog) | [Discord](https:\u002F\u002Fdiscord.gg\u002FmJeNZYRz4m) | [Twitter](https:\u002F\u002Ftwitter.com\u002Fneum_ai)\n  \n  \u003Ca href=\"https:\u002F\u002Fwww.ycombinator.com\u002Fcompanies\u002Fneum-ai\">\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FNeumTry_NeumAI_readme_64e8b3dee6f7.png\"\u002F>\u003C\u002Fa> \n  \u003Ca href=\"https:\u002F\u002Fpypi.org\u002Fproject\u002Fneumai\u002F\">\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fpypi\u002Fv\u002Fneumai\" alt=\"PyPI\">\n  \u003C\u002Fa>\n\u003C\u002Fdiv>\n\n![Neum AI 主页图](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FNeumTry_NeumAI_readme_0ec7eeb1b4a4.png)\n\n**[Neum AI](https:\u002F\u002Fneum.ai) 是一个数据平台，旨在帮助开发者通过检索增强生成（RAG）技术，利用其数据为大型语言模型提供上下文信息。** 这包括从现有数据源（如文档存储和 NoSQL 数据库）中提取数据、将内容处理为向量嵌入，并将这些向量嵌入导入向量数据库以进行相似性搜索。\n\n它为您提供了一套全面的 RAG 解决方案，能够随您的应用扩展，并减少在集成数据连接器、嵌入模型和向量数据库等服务上所花费的时间。\n\n## 功能\n\n- 🏭 **高吞吐量分布式架构**，可处理数十亿条数据。支持高度并行化，以优化嵌入生成和数据摄取。\n- 🧱 **内置数据连接器**，支持常见数据源、嵌入服务和向量存储。\n- 🔄 **实时数据同步**，确保您的数据始终是最新的。\n- ♻ **可定制的数据预处理**，包括加载、分块和选择。\n- 🤝 **一体化的数据管理**，支持结合元数据的混合检索。Neum AI 会自动增强并跟踪元数据，从而提供丰富的检索体验。\n\n## 联系我们\n\n您可以通过电子邮件 ([founders@tryneum.com](mailto:founders@tryneum.com))、[Discord](https:\u002F\u002Fdiscord.gg\u002FmJeNZYRz4m) 或 [预约通话](https:\u002F\u002Fcalendly.com\u002Fneum-ai\u002Fneum-ai-demo?month=2023-12) 联系我们的团队。\n\n## 开始使用\n\n### Neum AI 云\n\n立即在 [dashboard.neum.ai](https:\u002F\u002Fdashboard.neum.ai) 注册。请参阅我们的 [快速入门指南](https:\u002F\u002Fdocs.neum.ai\u002Fget-started\u002Fquickstart)，开始使用。\n\nNeum AI 云支持大规模分布式架构，可对数百万份文档进行向量嵌入处理。有关完整功能集，请参阅：[云与本地](https:\u002F\u002Fneumai.mintlify.app\u002Fget-started\u002Fcloud-vs-local)\n\n### 本地开发\n\n安装 [`neumai`](https:\u002F\u002Fpypi.org\u002Fproject\u002Fneumai\u002F) 包：\n\n```bash\npip install neumai\n```\n\n要创建您的第一个数据管道，请访问我们的 [快速入门指南](https:\u002F\u002Fdocs.neum.ai\u002Fget-started\u002Fquickstart)。\n\n从高层次来看，一个管道由一个或多个数据源、一个嵌入连接器（用于将内容向量化）以及一个目标连接器（用于存储这些向量）组成。以下代码片段将展示如何构建并运行一个管道：\n\u003Cdetails open>\u003Csummary>\n  \n  ### 创建并运行一个管道\n  \u003C\u002Fsummary>\n  \n  ```python\n  \n  from neumai.DataConnectors.WebsiteConnector import WebsiteConnector\n  from neumai.Shared.Selector import Selector\n  from neumai.Loaders.HTMLLoader import HTMLLoader\n  from neumai.Chunkers.RecursiveChunker import RecursiveChunker\n  from neumai.Sources.SourceConnector import SourceConnector\n  from neumai.EmbedConnectors import OpenAIEmbed\n  from neumai.SinkConnectors import WeaviateSink\n  from neumai.Pipelines import Pipeline\n\n  website_connector =  WebsiteConnector(\n      url = \"https:\u002F\u002Fwww.neum.ai\u002Fpost\u002Fretrieval-augmented-generation-at-scale\",\n      selector = Selector(\n          to_metadata=['url']\n      )\n  )\n  source = SourceConnector(\n      data_connector = website_connector, \n      loader = HTMLLoader(), \n      chunker = RecursiveChunker()\n  )\n\n  openai_embed = OpenAIEmbed(\n      api_key = \"\u003COPEN AI KEY>\",\n  )\n\n  weaviate_sink = WeaviateSink(\n      url = \"your-weaviate-url\",\n      api_key = \"your-api-key\",\n      class_name = \"your-class-name\",\n  )\n\n  pipeline = Pipeline(\n      sources=[source], \n      embed=openai_embed, \n      sink=weaviate_sink\n  )\n  pipeline.run()\n\n  results = pipeline.search(\n      query=\"What are the challenges with scaling RAG?\", \n      number_of_results=3\n  )\n\n  for result in results:\n      print(result.metadata)\n  ```\n\u003C\u002Fdetails>\n\n\u003Cdetails>\u003Csummary>\n\n  ### 创建并运行一个管道 - Postgres 连接器\n  \u003C\u002Fsummary>\n\n  ```python\n  \n  from neumai.DataConnectors.PostgresConnector import PostgresConnector\n  from neumai.Shared.Selector import Selector\n  from neumai.Loaders.JSONLoader import JSONLoader\n  from neumai.Chunkers.RecursiveChunker import RecursiveChunker\n  from neumai.Sources.SourceConnector import SourceConnector\n  from neumai.EmbedConnectors import OpenAIEmbed\n  from neumai.SinkConnectors import WeaviateSink\n  from neumai.Pipelines import Pipeline\n\n  website_connector =  PostgresConnector(\n      connection_string = 'postgres',\n      query = 'Select * from ...'\n  )\n  source = SourceConnector(\n      data_connector = website_connector, \n      loader = JSONLoader(\n          id_key='\u003Cyour id key of your jsons>',\n          selector=Selector(\n              to_embed=['property1_to_embed','property2_to_embed'],\n              to_metadata=['property3_to_include_in_metadata_in_vector']\n          )\n      ),\n      chunker = RecursiveChunker()\n  )\n\n  openai_embed = OpenAIEmbed(\n      api_key = \"\u003COPEN AI KEY>\",\n  )\n\n  weaviate_sink = WeaviateSink(\n      url = \"your-weaviate-url\",\n      api_key = \"your-api-key\",\n      class_name = \"your-class-name\",\n  )\n\n  pipeline = Pipeline(\n      sources=[source], \n      embed=openai_embed, \n      sink=weaviate_sink\n  )\n\n  pipeline.run()\n\n  results = pipeline.search(\n      query=\"...\", \n      number_of_results=3\n  )\n\n  for result in results:\n      print(result.metadata)\n  ```\n\u003C\u002Fdetails>\n\n\u003Cdetails>\u003Csummary>\n  \n  ### 将管道发布到 Neum Cloud\n  \u003C\u002Fsummary>\n  \n  ```python\n  from neumai.Client.NeumClient import NeumClient\n  client = NeumClient(\n      api_key='\u003Cyour neum api key, get it from https:\u002F\u002Fdashboard.neum.ai',\n  )\n  client.create_pipeline(pipeline=pipeline)\n  ```\n\u003C\u002Fdetails>\n\n### 自行托管\n\n如果您有兴趣将 Neum AI 部署到您自己的云环境中，请通过 [founders@tryneum.com](mailto:founders@tryneum.com) 联系我们。\n\n我们在 [GitHub](https:\u002F\u002Fgithub.com\u002FNeumTry\u002Fneum-at-scale) 上发布了一个示例后端架构，您可以将其作为起点。\n\n## 可用连接器\n如需查看最新列表，请访问我们的 [文档](https:\u002F\u002Fdocs.neum.ai\u002Fcomponents\u002FsourceConnector)。\n\n\u003Cdetails>\n\n### 源连接器\n1. Postgres\n2. 托管文件\n3. 网站\n4. S3\n5. Azure Blob\n6. Sharepoint\n7. Singlestore\n8. Supabase Storage\n\n### 嵌入连接器\n1. OpenAI 嵌入\n2. Azure OpenAI 嵌入\n\n### 目标连接器\n1. Supabase PostgreSQL\n2. Weaviate\n3. Qdrant\n4. Pinecone\n5. Singlestore\n\n\u003C\u002Fdetails>\n\n## 路线图\n我们的路线图会根据需求不断调整，如果您觉得有遗漏的功能，请随时提交问题或联系我们。\n\n\u003Cdetails>\n  \n连接器\n- [ ]  MySQL - 源\n- [ ]  GitHub - 源\n- [ ]  Google Drive - 源\n- [ ]  Hugging Face - 嵌入\n- [x]  LanceDB - 目标\n- [x]  Marqo - 目标\n- [ ]  Milvus - 目标\n- [ ]  Chroma - 目标\n\n搜索\n- [x]  检索反馈\n- [x]  过滤支持\n- [x]  统一的 Neum AI 过滤器\n- [ ]  智能路由（基于嵌入的分类）\n- [ ]  智能路由（基于 LLM 的分类）\n- [ ]  自查询检索（结合元数据属性生成）\n\n可扩展性\n- [x]  Langchain \u002F Llama Index 文档到 Neum 文档转换器\n- [ ]  自定义分块与加载\n\n实验性功能\n- [ ]  异步元数据增强\n- [ ]  聊天历史连接器\n- [ ]  结构化（SQL 和 GraphQL）搜索连接器\n\u003C\u002Fdetails>\n\n## Neum 工具\nNeum AI 的其他工具可以在这里找到：\n\n- [neumai-tools](https:\u002F\u002Fpypi.org\u002Fproject\u002Fneumai-tools\u002F)：包含用于在生成向量嵌入之前对数据进行加载和分块的预处理工具。","# Neum AI 快速上手指南\n\nNeum AI 是一个专为检索增强生成（RAG）打造的数据平台，帮助开发者从现有数据源（如文档存储、NoSQL）提取数据，处理为向量嵌入，并注入向量数据库以支持相似度搜索。\n\n## 环境准备\n\n*   **系统要求**：支持 Python 3.8+ 的操作系统（Linux, macOS, Windows）。\n*   **前置依赖**：\n    *   Python 包管理工具 `pip`。\n    *   有效的 OpenAI API Key（用于生成向量嵌入）。\n    *   一个可用的向量数据库实例（如 Weaviate, Qdrant, Pinecone 等）或使用 Neum AI Cloud。\n\n## 安装步骤\n\n使用 pip 安装核心库：\n\n```bash\npip install neumai\n```\n\n> **提示**：国内用户若下载缓慢，可使用清华或阿里镜像源加速安装：\n> ```bash\n> pip install neumai -i https:\u002F\u002Fpypi.tuna.tsinghua.edu.cn\u002Fsimple\n> ```\n\n如需使用额外的数据预处理工具，可安装辅助包：\n```bash\npip install neumai-tools\n```\n\n## 基本使用\n\n以下示例展示如何构建一个简单的 RAG 管道：从网站抓取内容，使用 OpenAI 进行向量化，并存入 Weaviate 向量数据库，最后执行搜索。\n\n### 1. 构建并运行管道\n\n```python\nfrom neumai.DataConnectors.WebsiteConnector import WebsiteConnector\nfrom neumai.Shared.Selector import Selector\nfrom neumai.Loaders.HTMLLoader import HTMLLoader\nfrom neumai.Chunkers.RecursiveChunker import RecursiveChunker\nfrom neumai.Sources.SourceConnector import SourceConnector\nfrom neumai.EmbedConnectors import OpenAIEmbed\nfrom neumai.SinkConnectors import WeaviateSink\nfrom neumai.Pipelines import Pipeline\n\n# 1. 配置数据源 (例如抓取网页)\nwebsite_connector = WebsiteConnector(\n    url=\"https:\u002F\u002Fwww.neum.ai\u002Fpost\u002Fretrieval-augmented-generation-at-scale\",\n    selector=Selector(to_metadata=['url'])\n)\n\nsource = SourceConnector(\n    data_connector=website_connector, \n    loader=HTMLLoader(), \n    chunker=RecursiveChunker()\n)\n\n# 2. 配置嵌入模型 (OpenAI)\nopenai_embed = OpenAIEmbed(\n    api_key=\"\u003COPEN AI KEY>\", # 请替换为你的真实 Key\n)\n\n# 3. 配置向量数据库接收端 (Weaviate)\nweaviate_sink = WeaviateSink(\n    url=\"your-weaviate-url\",       # 替换为你的 Weaviate 地址\n    api_key=\"your-api-key\",        # 替换为你的 API Key\n    class_name=\"your-class-name\",  # 替换为你的类名\n)\n\n# 4. 创建并运行管道\npipeline = Pipeline(\n    sources=[source], \n    embed=openai_embed, \n    sink=weaviate_sink\n)\n\npipeline.run()\n```\n\n### 2. 执行相似度搜索\n\n数据入库后，可直接通过管道对象进行查询：\n\n```python\nresults = pipeline.search(\n    query=\"What are the challenges with scaling RAG?\", \n    number_of_results=3\n)\n\nfor result in results:\n    print(result.metadata)\n```\n\n### 进阶场景：连接 PostgreSQL\n\n若需从数据库读取数据，可切换 `SourceConnector`：\n\n```python\nfrom neumai.DataConnectors.PostgresConnector import PostgresConnector\nfrom neumai.Loaders.JSONLoader import JSONLoader\n# ... 其他导入保持不变\n\npostgres_connector = PostgresConnector(\n    connection_string='postgres', # 替换为实际连接字符串\n    query='Select * from your_table'\n)\n\nsource = SourceConnector(\n    data_connector=postgres_connector, \n    loader=JSONLoader(\n        id_key='\u003Cyour id key>',\n        selector=Selector(\n            to_embed=['property1_to_embed', 'property2_to_embed'],\n            to_metadata=['property3_to_include']\n        )\n    ),\n    chunker=RecursiveChunker()\n)\n\n# 后续嵌入和 Sink 配置同上，创建 Pipeline 并运行即可\n```\n\n### 部署到 Neum Cloud\n\n若希望利用云端分布式架构处理大规模数据，可将本地定义的管道发布至 Neum Cloud：\n\n```python\nfrom neumai.Client.NeumClient import NeumClient\n\nclient = NeumClient(\n    api_key='\u003Cyour neum api key>', # 从 dashboard.neum.ai 获取\n)\n\nclient.create_pipeline(pipeline=pipeline)\n```","某电商初创团队正致力于构建一个基于 RAG 的智能客服系统，需要实时同步海量商品文档至向量数据库以支持精准问答。\n\n### 没有 NeumAI 时\n- **开发周期漫长**：工程师需手动编写代码分别对接 MongoDB 商品库、调用 OpenAI 嵌入接口并写入 Weaviate，重复造轮子耗时数周。\n- **数据更新滞后**：缺乏自动同步机制，新上架商品或价格变动无法实时反映在知识库中，导致客服回答过时信息。\n- **扩展能力受限**：面对百万级商品数据，自建的单线程处理脚本经常内存溢出，难以通过并行计算优化吞吐量。\n- **元数据管理混乱**：在切片和向量化过程中容易丢失商品分类、品牌等关键元数据，严重影响混合检索的准确率。\n\n### 使用 NeumAI 后\n- **快速落地管道**：利用 NeumAI 内置的 NoSQL 连接器、OpenAI 嵌入模块及 Weaviate 接收器，仅需几行 Python 代码即可在半天内搭建完整数据流。\n- **实时同步保障**：借助 NeumAI 的实时同步功能，一旦后台商品库更新，向量索引即刻自动刷新，确保客服始终基于最新数据回答。\n- **亿级数据处理**：依托 NeumAI 的高吞吐分布式架构，系统轻松并行处理千万级商品文档，生成与入库效率提升数十倍。\n- **智能元数据增强**：NeumAI 自动在加载、切片阶段保留并增强元数据，完美支持“按品牌 + 语义”的混合检索，大幅提升回答相关性。\n\nNeumAI 将原本繁琐脆弱的多服务集成工作转化为标准化、可扩展的自动化流程，让团队能专注于核心业务逻辑而非数据基建。","https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FNeumTry_NeumAI_0ec7eeb1.png","NeumTry","Neum AI","https:\u002F\u002Foss.gittoolsai.com\u002Favatars\u002FNeumTry_1273da96.png","Also VetRec",null,"founders@tryneum.com","neum_ai","tryneum.com","https:\u002F\u002Fgithub.com\u002FNeumTry",[83],{"name":84,"color":85,"percentage":86},"Python","#3572A5",100,868,50,"2026-04-16T07:56:24","Apache-2.0","未说明",{"notes":93,"python":91,"dependencies":94},"该工具主要作为 Python 库（pip install neumai）使用，支持本地开发或云端部署。本地运行时需自行配置外部依赖服务，包括嵌入模型提供商（如 OpenAI API）和向量数据库（如 Weaviate, Qdrant, Pinecone 等）。README 中未明确指定具体的操作系统、GPU、内存或 Python 版本要求，通常意味着它依赖于宿主环境及所连接的外部服务的环境配置。若需自托管大规模分布式架构，可参考其 GitHub 上的后端架构示例。",[95,96],"neumai","neumai-tools",[35,16,98,15,13,14],"其他",[100,101,102,103,104,105,106,107,108,109,110,111,112,113,114,115,116],"ai","data","embeddings","etl","llm","vector-database","chatgpt","data-engineering","database","pipeline","python","retrieval","vectors","llmops","mlops","ops","rag","2026-03-27T02:49:30.150509","2026-04-18T14:25:45.645055",[120,125,130,135,140,145],{"id":121,"question_zh":122,"answer_zh":123,"source_url":124},40056,"遇到 'ModuleNotFoundError: No module named neumai_tools' 错误怎么办？","这是一个已知的依赖问题。短期解决方法是手动安装缺失的包：运行 `pip install neumai-tools`。该依赖要求已在后续版本（0.0.34 及更高）中移除或修复，建议升级库版本以彻底解决。","https:\u002F\u002Fgithub.com\u002FNeumTry\u002FNeumAI\u002Fissues\u002F30",{"id":126,"question_zh":127,"answer_zh":128,"source_url":129},40057,"如何在搜索方法中正确传递过滤条件（filters）？","用户应传递字典列表（`List[dict]`），而不是直接实例化 `FilterCondition` 对象。由于 `FilterCondition` 是基于 Pydantic 的模型，你可以直接使用解包语法将字典转换为对象，例如：`FilterCondition(**your_dict)`。此外，参数命名已统一为 `filters`，请确保代码中使用一致的参数名。","https:\u002F\u002Fgithub.com\u002FNeumTry\u002FNeumAI\u002Fissues\u002F50",{"id":131,"question_zh":132,"answer_zh":133,"source_url":134},40058,"是否支持范围查询（如大于、小于）或不等于条件的过滤？","是的，项目已更新以支持更复杂的过滤条件。现在可以通过字符串形式提供过滤条件（例如 `\"field1 \u003C= value1, field2 != value2\"`），系统会将其解析为 `FilterCondition` 对象。该功能已在所有现有的 Sink 连接器中实现，支持操作符和嵌套条件。","https:\u002F\u002Fgithub.com\u002FNeumTry\u002FNeumAI\u002Fissues\u002F39",{"id":136,"question_zh":137,"answer_zh":138,"source_url":139},40059,"LanceDBSink 搜索时出现 'unexpected keyword argument filters' 错误如何解决？","该问题已在最新版本中修复。如果遇到此错误，请确保你使用的是包含修复补丁的版本。官方文档已更新，请参考 LanceDBSink 的最新用法指南：https:\u002F\u002Fdocs.neum.ai\u002Fcomponents\u002Fsink-connectors\u002FLanceDBSink。","https:\u002F\u002Fgithub.com\u002FNeumTry\u002FNeumAI\u002Fissues\u002F52",{"id":141,"question_zh":142,"answer_zh":143,"source_url":144},40060,"HuggingFace 嵌入连接器的参数验证似乎无效，如何修复？","此前验证逻辑仅默认返回 true，现已通过 PR #62 修复。更新后的版本会真正连接客户端以验证参数有效性。请升级到包含该修复的最新版本以确保参数校验正常工作。","https:\u002F\u002Fgithub.com\u002FNeumTry\u002FNeumAI\u002Fissues\u002F58",{"id":146,"question_zh":147,"answer_zh":148,"source_url":134},40061,"如何在不同数据库后端之间统一过滤条件的映射？","项目引入了统一的 `FilterCondition` 类来处理过滤逻辑。用户提供的过滤条件（无论是字典还是字符串）会被解析并映射到底层数据库特定的过滤选项上。目前实现了对多种操作符的支持，并计划根据反馈增加更多嵌套条件的支持。",[150,155,160,165,170,175,180,184],{"id":151,"version":152,"summary_zh":153,"released_at":154},323571,"neumai-0.0.40","功能：HuggingFace 嵌入 @sunilkumardash9  \n修复：WeaviateSink 搜索过滤器问题 @ashi-agrawal  \n修复：LanceDB 配置已明确公开 @sky-2002","2024-01-07T17:38:43",{"id":156,"version":157,"summary_zh":158,"released_at":159},323572,"neumai-0.0.39","修复：标准化搜索中 `filters` 的使用 #50 @ddematheu\n修复：LanceDB 问题 #52 @ddematheu\n","2024-01-02T18:54:04",{"id":161,"version":162,"summary_zh":163,"released_at":164},323573,"neumai-0.038","针对向量数据库的过滤条件 @sky-2002 @ddematheu  \n在 Marqo、Weaviate、Pinecone、Qdrant、Supabase 和 SingleStore 中的支持 @sky-2002 @ddematheu  \nLanceDB 问题修复 @PrashantDixit0  \n搜索结果修复：使得分成为可选项 @ddematheu ","2023-12-29T16:52:52",{"id":166,"version":167,"summary_zh":168,"released_at":169},323574,"neumai-0.0.37","特性：@sky-2002 提供 Marqo 数据库支持  \n特性：@sky-2002 提供 Lance 数据库支持","2023-12-19T18:30:49",{"id":171,"version":172,"summary_zh":173,"released_at":174},323575,"neumai-0.0.36","特性：客户端更新以获取检索结果\n\n特性：为对象添加处理时间\n\n修复：修复管道运行问题\n\n修复：修复 Weaviate 类名问题","2023-12-13T04:04:20",{"id":176,"version":177,"summary_zh":178,"released_at":179},323576,"neuami-0.0.35","功能：更新以支持实时 S3  \n功能：支持向量删除  \n功能：Neum 客户端新增更多获取方法  \n修复：网站和文件连接器的响应检查","2023-12-12T21:47:53",{"id":181,"version":96,"summary_zh":182,"released_at":183},323577,"将 neumai 版本升级至 0.0.34","2023-12-11T22:54:08",{"id":185,"version":95,"summary_zh":186,"released_at":187},323578,"修复：网站稳定性增强  \n修复：neum-tools 依赖问题  \n修复：S3 前缀问题  \n特性：为 Weaviate、Pinecone 和 Supabase 的搜索添加元数据过滤器  \n特性：客户端更新，新增多个 API 端点","2023-12-11T22:42:53"]