[{"data":1,"prerenderedAt":9250},["ShallowReactive",2],{"header-counts":3,"footer-counts":6,"tool-\u002Ftools\u002Fagent\u002Fplatform\u002Fcrewai":8,"cat-rank-agent-platform":613,"tool-related-agent\u002Fplatform\u002Fcrewai":7903,"tool-reviews-agent\u002Fplatform\u002Fcrewai":7904,"tool-alts-agent\u002Fplatform\u002Fcrewai":7905},{"tools":4,"reviews":5},98,27,{"tools":4,"reviews":5,"playbooks":7,"news":5},22,{"id":9,"title":10,"alternatives":11,"api_compatible":15,"body":16,"category":577,"chinese_friendly":566,"cover":578,"description":579,"domestic":580,"extension":581,"faq":15,"free":580,"github":558,"languages":582,"lastVerified":584,"meta":585,"models":15,"navigation":586,"notSuitable":15,"opensource":586,"path":587,"pillar":588,"platforms":589,"priceTable":15,"pricing":592,"published":593,"relatedPlaybooks":15,"relatedReviews":15,"score":594,"self_host":580,"seo":597,"seoTitle":598,"slug":599,"sources":600,"stem":603,"suitable":15,"tagline":604,"tags":605,"updated":584,"verdict":611,"website":550,"__hash__":612},"tools\u002Ftools\u002Fagent\u002Fplatform\u002Fcrewai.md","CrewAI",[12,13,14],"agent\u002Fplatform\u002Flangflow","agent\u002Fplatform\u002Fautogen","agent\u002Fplatform\u002Fn8n",null,{"type":17,"value":18,"toc":561},"minimark",[19,24,28,31,34,93,96,150,156,160,168,191,196,219,222,264,267,403,406,454,458,490,494,500,506,512,518,521,536,539,544],[20,21,23],"h2",{"id":22},"tldr","TL;DR",[25,26,27],"p",{},"CrewAI 是开源多 Agent 协作框架（MIT 协议），用 Python 代码定义 Agent 角色（Role）、目标（Goal）、工具（Tools），组合成 Crew 执行任务序列。核心概念清晰——Agent 负责做事、Task 定义做什么、Crew 编排怎么协作。支持顺序 \u002F 层级 \u002F 自定义流程，内置 50+ 工具集成。CrewAI Enterprise 提供云端托管 + 可视化监控。",[25,29,30],{},"适合：需要多 Agent 自动化工作流的开发者、Python 技术栈团队、快速原型验证多 Agent 方案。不适合：非技术用户（用 Dify）、需要极复杂条件分支流程（用 LangGraph）、需要 GUI 可视化编排（用 Flowise）。",[20,32,33],{"id":33},"核心能力",[35,36,37,45,51,57,63,69,75,81,87],"ul",{},[38,39,40,44],"li",{},[41,42,43],"strong",{},"角色定义","：Agent = Role + Goal + Backstory + Tools，角色人设驱动行为",[38,46,47,50],{},[41,48,49],{},"任务编排","：Task 定义具体任务 + 期望输出 + 分配 Agent",[38,52,53,56],{},[41,54,55],{},"Crew 编排","：顺序执行 \u002F 层级管理 \u002F 自定义流程三种模式",[38,58,59,62],{},[41,60,61],{},"工具集成","：内置 SerperDev \u002F Firecrawl \u002F FileRead \u002F WebScraper 等 50+ 工具",[38,64,65,68],{},[41,66,67],{},"流程控制","：支持任务间依赖、条件路由、输出传递",[38,70,71,74],{},[41,72,73],{},"记忆系统","：Short-term \u002F Long-term \u002F Entity Memory，Agent 可跨任务记忆",[38,76,77,80],{},[41,78,79],{},"多模型支持","：OpenAI \u002F Claude \u002F Gemini \u002F Ollama \u002F 任意 LiteLLM 兼容模型",[38,82,83,86],{},[41,84,85],{},"CrewAI Enterprise","：云端托管 + 可视化 Crew 监控 + 团队协作",[38,88,89,92],{},[41,90,91],{},"输出结构化","：支持 Pydantic 模型定义输出格式",[20,94,95],{"id":95},"价格",[97,98,99,114],"table",{},[100,101,102],"thead",{},[103,104,105,109,111],"tr",{},[106,107,108],"th",{},"方案",[106,110,95],{},[106,112,113],{},"核心功能",[115,116,117,129,139],"tbody",{},[103,118,119,123,126],{},[120,121,122],"td",{},"开源版",[120,124,125],{},"$0",[120,127,128],{},"完整框架，MIT 协议，本地运行",[103,130,131,133,136],{},[120,132,85],{},[120,134,135],{},"$49\u002F月起",[120,137,138],{},"云端托管 + 可视化监控 + API",[103,140,141,144,147],{},[120,142,143],{},"Enterprise+",[120,145,146],{},"联系销售",[120,148,149],{},"SSO \u002F 私有部署 \u002F 专属支持",[151,152,153],"blockquote",{},[25,154,155],{},"价格信息基于 2026-07 官网，可能调整。",[20,157,159],{"id":158},"体验与评测资料整理","体验与评测（资料整理）",[151,161,162],{},[25,163,164,165],{},"说明：本节基于官方文档与公开评测整理，非本站独立实测环境，具体数据请以官方为准。\n",[41,166,167],{},"亮点：",[35,169,170,173,176,179,182,185,188],{},[38,171,172],{},"API 设计非常直观，Agent + Task + Crew 三件套 10 分钟上手",[38,174,175],{},"角色人设（Backstory）确实影响 Agent 行为，写好 backstory 效果提升明显",[38,177,178],{},"层级模式（hierarchical）下 Manager Agent 自动分配任务，适合复杂场景",[38,180,181],{},"内置工具丰富，SerperDev 搜索 + Firecrawl 爬虫开箱即用",[38,183,184],{},"Memory 系统让 Agent 跨任务保持上下文，长流程不丢信息",[38,186,187],{},"Pydantic 结构化输出对后续处理非常友好",[38,189,190],{},"CrewAI Enterprise 的可视化监控能看到每个 Agent 的思考过程",[25,192,193],{},[41,194,195],{},"踩坑：",[35,197,198,201,204,207,210,213,216],{},[38,199,200],{},"Agent 偶尔\"不听话\"——偏离角色设定、重复执行、跳过任务",[38,202,203],{},"复杂流程的调试困难，错误信息不够清晰",[38,205,206],{},"Token 消耗不小——多 Agent + 多轮对话 + Memory 存储",[38,208,209],{},"开源版无 GUI，全靠日志调试，Enterprise 版才有可视化",[38,211,212],{},"版本迭代快，API 偶有 breaking changes",[38,214,215],{},"层级模式的 Manager Agent 判断不稳定，有时分配不合理",[38,217,218],{},"中文 system prompt 效果不如英文，建议角色定义用英文",[20,220,221],{"id":221},"上手",[223,224,225,232,239,245,251,257],"ol",{},[38,226,227,231],{},[228,229,230],"code",{},"pip install crewai crewai-tools","（建议用 uv 管理虚拟环境）",[38,233,234,235,238],{},"配置 LLM：设置 ",[228,236,237],{},"OPENAI_API_KEY"," 或在代码中指定 model",[38,240,241,242],{},"定义 Agent：",[228,243,244],{},"Agent(role=\"研究员\", goal=\"搜集信息\", tools=[search_tool])",[38,246,247,248],{},"定义 Task：",[228,249,250],{},"Task(description=\"调研XX趋势\", agent=researcher, expected_output=\"报告\")",[38,252,253,254],{},"组建 Crew：",[228,255,256],{},"Crew(agents=[researcher, writer], tasks=[task1, task2], process=Process.sequential)",[38,258,259,260,263],{},"启动：",[228,261,262],{},"result = crew.kickoff()","，查看结果 + 日志",[20,265,266],{"id":266},"对比",[97,268,269,287],{},[100,270,271],{},[103,272,273,276,278,281,284],{},[106,274,275],{},"维度",[106,277,10],{},[106,279,280],{},"AutoGen",[106,282,283],{},"LangGraph",[106,285,286],{},"Dify",[115,288,289,305,322,339,355,371,386],{},[103,290,291,294,297,300,302],{},[120,292,293],{},"上手难度",[120,295,296],{},"低",[120,298,299],{},"高",[120,301,299],{},[120,303,304],{},"极低",[103,306,307,310,313,316,319],{},[120,308,309],{},"API 设计",[120,311,312],{},"优雅直观",[120,314,315],{},"底层灵活",[120,317,318],{},"图模型",[120,320,321],{},"可视化",[103,323,324,327,330,333,336],{},[120,325,326],{},"多 Agent",[120,328,329],{},"✅ Crew 角色",[120,331,332],{},"✅ Group Chat",[120,334,335],{},"✅ 图编排",[120,337,338],{},"✅ 工作流",[103,340,341,344,347,350,352],{},[120,342,343],{},"代码执行",[120,345,346],{},"需自定义",[120,348,349],{},"✅ Docker 沙箱",[120,351,346],{},[120,353,354],{},"沙箱",[103,356,357,359,362,365,368],{},[120,358,73],{},[120,360,361],{},"✅ 内置",[120,363,364],{},"有限",[120,366,367],{},"需自建",[120,369,370],{},"✅",[103,372,373,376,379,382,384],{},[120,374,375],{},"GUI",[120,377,378],{},"Enterprise 版",[120,380,381],{},"❌",[120,383,381],{},[120,385,370],{},[103,387,388,391,394,397,400],{},[120,389,390],{},"适合场景",[120,392,393],{},"业务自动化",[120,395,396],{},"研究",[120,398,399],{},"精确流程",[120,401,402],{},"应用构建",[20,404,405],{"id":405},"避坑",[35,407,408,414,420,426,432,438,448],{},[38,409,410,413],{},[41,411,412],{},"Backstory 认真写","：角色人设直接影响 Agent 行为质量，模糊描述 = 模糊行为",[38,415,416,419],{},[41,417,418],{},"expected_output 必填","：不定义预期输出，Agent 容易跑偏",[38,421,422,425],{},[41,423,424],{},"控制 Agent 数量","：3-5 个 Agent 最佳，超过 8 个协调成本急升",[38,427,428,431],{},[41,429,430],{},"Memory 按需开启","：Long-term Memory 会累积 token 消耗，简单任务关掉",[38,433,434,437],{},[41,435,436],{},"调试用 verbose=True","：开启详细日志看 Agent 思考过程，定位问题",[38,439,440,443,444,447],{},[41,441,442],{},"版本锁定","：",[228,445,446],{},"pip install crewai==x.x.x","，迭代快别用 latest",[38,449,450,453],{},[41,451,452],{},"层级模式慎用","：Manager Agent 不稳定，简单场景用 sequential 更可靠",[20,455,457],{"id":456},"适合-不适合","适合 \u002F 不适合",[35,459,460,463,466,469,472,475,478,481,484,487],{},[38,461,462],{},"✅ Python 开发者快速构建多 Agent 工作流",[38,464,465],{},"✅ 内容生产流水线（调研 → 写作 → 审核）",[38,467,468],{},"✅ 自动化研究 \u002F 数据收集 \u002F 报告生成",[38,470,471],{},"✅ 需要角色分工的协作场景",[38,473,474],{},"✅ 快速原型验证多 Agent 方案",[38,476,477],{},"❌ 非技术用户（用 Dify \u002F Flowise）",[38,479,480],{},"❌ 需要极复杂条件分支流程（用 LangGraph）",[38,482,483],{},"❌ 需要精细控制 Agent 对话轮次（用 AutoGen）",[38,485,486],{},"❌ 需要免费 GUI 可视化监控（开源版无 GUI）",[38,488,489],{},"❌ 预算敏感的高频调用场景（多 Agent token 消耗大）",[20,491,493],{"id":492},"faq","FAQ",[25,495,496,499],{},[41,497,498],{},"Q: CrewAI 和 AutoGen 怎么选？","\nA: CrewAI 上手更快，Agent + Task + Crew 概念直观，适合业务自动化和快速原型。AutoGen 更底层灵活，Group Chat + 代码执行 + 事件驱动适合研究和复杂协作。做产品选 CrewAI，做研究选 AutoGen。",[25,501,502,505],{},[41,503,504],{},"Q: 开源版和 Enterprise 版差别大吗？","\nA: 开源版框架功能完整，能跑所有 Agent \u002F Task \u002F Crew。Enterprise 版主要多了云端托管（免运维）、可视化监控（看 Agent 思考过程）、团队协作和 API 服务。如果只是本地跑 Agent 工作流，开源版够用。",[25,507,508,511],{},[41,509,510],{},"Q: 可以接入本地模型吗？","\nA: 可以。CrewAI 基于 LiteLLM，支持 Ollama \u002F vLLM \u002F LM Studio 等本地模型。但本地模型能力有限，角色扮演和工具调用效果可能不如 GPT-4 \u002F Claude。建议开发用便宜模型，生产用高级模型。",[25,513,514,517],{},[41,515,516],{},"Q: Agent 总是跑偏怎么办？","\nA: 三步排查：1）检查 Backstory 是否足够具体；2）确认 expected_output 定义清晰；3）开启 verbose=True 看思考过程定位偏移点。复杂任务拆成更小的 Task，每个 Task 目标单一明确。",[20,519,520],{"id":520},"相关阅读",[25,522,523,527,528,527,532],{},[524,525,280],"a",{"href":526},"\u002Fagent\u002Fplatform\u002Fautogen.html"," · ",[524,529,531],{"href":530},"\u002Fagent\u002Fplatform\u002Fflowise.html","Flowise",[524,533,535],{"href":534},"\u002Fcoding\u002Fapi\u002Flangfuse.html","Langfuse",[20,537,538],{"id":538},"来源",[151,540,541],{},[25,542,543],{},"本文的价格、版本号与性能数据均参考以下官方渠道整理，可能随时间变动，请以官方实时信息为准。",[35,545,546,554],{},[38,547,548],{},[524,549,553],{"href":550,"rel":551},"https:\u002F\u002Fcrewai.com",[552],"nofollow","官网",[38,555,556],{},[524,557,560],{"href":558,"rel":559},"https:\u002F\u002Fgithub.com\u002FcrewAIInc\u002FcrewAI",[552],"GitHub",{"title":562,"searchDepth":563,"depth":563,"links":564},"",3,[565,567,568,569,570,571,572,573,574,575,576],{"id":22,"depth":566,"text":23},2,{"id":33,"depth":566,"text":33},{"id":95,"depth":566,"text":95},{"id":158,"depth":566,"text":159},{"id":221,"depth":566,"text":221},{"id":266,"depth":566,"text":266},{"id":405,"depth":566,"text":405},{"id":456,"depth":566,"text":457},{"id":492,"depth":566,"text":493},{"id":520,"depth":566,"text":520},{"id":538,"depth":566,"text":538},"platform","\u002Fimg\u002Ftools\u002Fcrewai.webp","CrewAI 真实评测：开源多 Agent 协作框架（MIT 协议），用 Python 代码定义 Agent 角色、任务和协作流程。支持角色分工、任务编排、工具集成，适合需要构建多 Agent 自动化工作流的开发者和企业团队。",false,"md",[583],"en","2026-07-30",{},true,"\u002Ftools\u002Fagent\u002Fplatform\u002Fcrewai","agent",[590,591],"linux","docker","Free \u002F 开源（MIT）\u002F Enterprise","2026-07-05",{"power":595,"ux":563,"price":596,"cn_support":566,"stability":563},4,5,{"title":10,"description":579},"CrewAI - 多 Agent 协作框架评测与使用 | AIHO","agent\u002Fplatform\u002Fcrewai",[601,602],{"title":553,"url":550},{"title":560,"url":558},"tools\u002Fagent\u002Fplatform\u002Fcrewai","多 Agent 协作框架，Python 代码定义角色和任务",[606,607,608,609,610],"agent-platform","multi-agent","framework","python","opensource","需要用 Python 快速构建多 Agent 自动化工作流的开发者首选，角色 + 任务 + Crew 的概念直观易学、API 设计优雅，但 Agent 对话可控性和稳定性不如 AutoGen，复杂流程编排需配合 LangGraph。","_O4h6a46IiWmQgKW4ommaTGIrJPA8tvgaA2kCNf0aa0",[614,1117,1589,2501,2897,3937,5011,5531,6152,6856,7358],{"id":615,"title":616,"alternatives":617,"api_compatible":15,"body":620,"category":577,"chinese_friendly":563,"cover":1093,"description":1094,"domestic":580,"extension":581,"faq":15,"free":580,"github":1078,"languages":1095,"lastVerified":584,"meta":1096,"models":15,"navigation":586,"notSuitable":15,"opensource":586,"path":1097,"pillar":588,"platforms":1098,"priceTable":15,"pricing":1101,"published":593,"relatedPlaybooks":15,"relatedReviews":15,"score":1102,"self_host":580,"seo":1103,"seoTitle":1104,"slug":1105,"sources":1106,"stem":1109,"suitable":15,"tagline":1110,"tags":1111,"updated":584,"verdict":1115,"website":1072,"__hash__":1116},"tools\u002Ftools\u002Fagent\u002Fplatform\u002Fanythingllm.md","AnythingLLM",[618,619,12],"agent\u002Fplatform\u002Fdify","agent\u002Fplatform\u002Ffastgpt",{"type":17,"value":621,"toc":1080},[622,624,627,630,632,688,690,734,738,740,746,770,774,794,796,824,826,946,948,989,991,1020,1022,1028,1034,1040,1046,1048,1060,1062,1066],[20,623,23],{"id":22},[25,625,626],{},"AnythingLLM 是 Mintplex Labs 出品的开源私有部署 LLM 平台，MIT 协议，主打\"一站式 RAG 知识库 + Agent + 多用户权限管理\"。桌面应用 \u002F Docker 双部署模式，接入 OpenAI \u002F Claude \u002F Ollama \u002F Azure 等任意模型，Workspaces 隔离不同知识库，内置向量数据库。适合需要私有化部署 AI 知识库且不写代码的团队。",[25,628,629],{},"适合：企业内网知识库、团队共享 AI 助手、需要多用户权限控制的私有化场景。不适合：需要复杂 Agent 编排（用 Dify \u002F Langflow）、需要极致 RAG 召回精度（用 RAGFlow \u002F FastGPT）、需要大规模并发生产服务。",[20,631,33],{"id":33},[35,633,634,640,646,652,658,664,670,676,682],{},[38,635,636,639],{},[41,637,638],{},"私有化部署","：Docker \u002F 桌面应用（Win\u002FMac\u002FLinux），数据完全在内网",[38,641,642,645],{},[41,643,644],{},"Workspaces 知识库隔离","：不同工作区独立向量库 + 文档 + 对话历史",[38,647,648,651],{},[41,649,650],{},"多用户权限管理","：管理员 \u002F 用户 \u002F 多工作区角色分配，适合团队使用",[38,653,654,657],{},[41,655,656],{},"任意模型接入","：OpenAI \u002F Claude \u002F Azure \u002F Ollama \u002F LM Studio \u002F 本地模型",[38,659,660,663],{},[41,661,662],{},"多向量数据库","：内置 LanceDB，可选 Chroma \u002F Pinecone \u002F Weaviate \u002F Qdrant",[38,665,666,669],{},[41,667,668],{},"文档处理","：PDF \u002F Word \u002F Excel \u002F TXT \u002F Markdown \u002F 网页链接，自动切片 + 向量化",[38,671,672,675],{},[41,673,674],{},"Agent 能力","：内置 Web 搜索 \u002F RAG 搜索 \u002F SQL 查询等工具调用",[38,677,678,681],{},[41,679,680],{},"嵌入向量","：支持自定义 embedding 模型，兼容 OpenAI \u002F 本地嵌入",[38,683,684,687],{},[41,685,686],{},"API 接口","：提供完整 REST API，可集成到外部系统",[20,689,95],{"id":95},[97,691,692,702],{},[100,693,694],{},[103,695,696,698,700],{},[106,697,108],{},[106,699,95],{},[106,701,113],{},[115,703,704,713,724],{},[103,705,706,708,710],{},[120,707,122],{},[120,709,125],{},[120,711,712],{},"完整功能，MIT 协议，自托管",[103,714,715,718,721],{},[120,716,717],{},"Cloud",[120,719,720],{},"$30\u002F月起",[120,722,723],{},"托管服务，免去运维，含团队协作",[103,725,726,729,731],{},[120,727,728],{},"Enterprise",[120,730,146],{},[120,732,733],{},"SSO \u002F 审计日志 \u002F 私有部署支持",[151,735,736],{},[25,737,155],{},[20,739,159],{"id":158},[151,741,742],{},[25,743,164,744],{},[41,745,167],{},[35,747,748,755,758,761,764,767],{},[38,749,750,751,754],{},"Docker 部署极快，一条 ",[228,752,753],{},"docker-compose up"," 起来就能用",[38,756,757],{},"Workspaces 隔离设计实用，不同部门知识库互不干扰",[38,759,760],{},"接 Ollama 本地模型完全离线运行，数据不出内网",[38,762,763],{},"桌面应用适合个人用户，安装即用零配置",[38,765,766],{},"文档上传后自动切片 + 向量化，问答效果在通用场景下可接受",[38,768,769],{},"多用户权限管理是开源 RAG 平台中少有的完整实现",[25,771,772],{},[41,773,195],{},[35,775,776,779,782,785,788,791],{},[38,777,778],{},"文档切片策略偏简单（固定长度），复杂表格 \u002F 图文混排召回效果一般",[38,780,781],{},"大文件（100MB+ PDF）处理偶尔超时，需调超时参数",[38,783,784],{},"Agent 能力有限，复杂工具链编排不如 Dify",[38,786,787],{},"向量库默认 LanceDB 在数据量大时查询变慢，建议切 Qdrant \u002F Chroma",[38,789,790],{},"UI 偶有卡顿，文档列表加载慢",[38,792,793],{},"中文文档的 OCR 需要额外配置，默认对扫描件支持有限",[20,795,221],{"id":221},[223,797,798,805,812,815,818,821],{},[38,799,800,801,804],{},"Docker 部署：",[228,802,803],{},"docker-compose up -d","（官方提供 docker-compose.yml）",[38,806,807,808,811],{},"首次访问 ",[228,809,810],{},"http:\u002F\u002Flocalhost:3001","，创建管理员账号",[38,813,814],{},"Settings → LLM Provider 配置模型（OpenAI API Key 或 Ollama 地址）",[38,816,817],{},"创建 Workspace → 上传文档（PDF\u002FWord\u002FTXT）",[38,819,820],{},"等待文档向量化完成，在 Chat 中开始问答",[38,822,823],{},"Settings → Users 添加团队成员并分配工作区权限",[20,825,266],{"id":266},[97,827,828,844],{},[100,829,830],{},[103,831,832,834,836,838,841],{},[106,833,275],{},[106,835,616],{},[106,837,286],{},[106,839,840],{},"FastGPT",[106,842,843],{},"Langflow",[115,845,846,860,874,887,903,917,930],{},[103,847,848,851,853,856,858],{},[120,849,850],{},"部署门槛",[120,852,304],{},[120,854,855],{},"中",[120,857,855],{},[120,859,855],{},[103,861,862,865,868,870,872],{},[120,863,864],{},"多用户权限",[120,866,867],{},"✅ 完整",[120,869,370],{},[120,871,370],{},[120,873,381],{},[103,875,876,879,881,883,885],{},[120,877,878],{},"RAG 精度",[120,880,855],{},[120,882,299],{},[120,884,299],{},[120,886,855],{},[103,888,889,892,895,898,900],{},[120,890,891],{},"Agent 编排",[120,893,894],{},"基础",[120,896,897],{},"强",[120,899,855],{},[120,901,902],{},"强（可视化）",[103,904,905,908,911,913,915],{},[120,906,907],{},"模型接入",[120,909,910],{},"丰富",[120,912,910],{},[120,914,910],{},[120,916,910],{},[103,918,919,922,924,926,928],{},[120,920,921],{},"桌面应用",[120,923,370],{},[120,925,381],{},[120,927,381],{},[120,929,381],{},[103,931,932,935,938,941,944],{},[120,933,934],{},"协议",[120,936,937],{},"MIT",[120,939,940],{},"Apache 2.0",[120,942,943],{},"FastGPT Open",[120,945,937],{},[20,947,405],{"id":405},[35,949,950,956,962,968,977,983],{},[38,951,952,955],{},[41,953,954],{},"切片策略默认偏简单","：对结构化文档（表格\u002F代码）效果差，可调 chunk size",[38,957,958,961],{},[41,959,960],{},"LanceDB 大数据量变慢","：文档超过 1 万条建议切 Qdrant 或 Chroma",[38,963,964,967],{},[41,965,966],{},"大文件超时","：调整 Docker 超时配置，或拆分文档上传",[38,969,970,973,974],{},[41,971,972],{},"Ollama 连接","：Docker 内访问宿主机 Ollama 需用 ",[228,975,976],{},"host.docker.internal",[38,978,979,982],{},[41,980,981],{},"embedding 模型选择","：中文场景建议用 bge-large-zh 而非默认 OpenAI embedding",[38,984,985,988],{},[41,986,987],{},"不要当生产级 Agent 平台用","：Agent 能力是辅助，复杂编排上 Dify",[20,990,457],{"id":456},[35,992,993,996,999,1002,1005,1008,1011,1014,1017],{},[38,994,995],{},"✅ 企业内网私有化 AI 知识库",[38,997,998],{},"✅ 团队共享 AI 助手 + 多用户权限管理",[38,1000,1001],{},"✅ 接 Ollama 完全离线运行",[38,1003,1004],{},"✅ 个人桌面端快速体验 RAG",[38,1006,1007],{},"✅ 需要快速验证 RAG 概念的原型项目",[38,1009,1010],{},"❌ 需要复杂 Agent 工作流编排（用 Dify \u002F Langflow）",[38,1012,1013],{},"❌ 需要极致 RAG 召回精度（用 RAGFlow \u002F FastGPT）",[38,1015,1016],{},"❌ 大规模并发生产服务（架构未做高可用）",[38,1018,1019],{},"❌ 需要深度文档解析（复杂表格\u002F公式\u002F扫描件）",[20,1021,493],{"id":492},[25,1023,1024,1027],{},[41,1025,1026],{},"Q: AnythingLLM 和 Dify 怎么选？","\nA: AnythingLLM 更轻量，部署快、有桌面应用、多用户权限开箱即用，适合快速搭建团队知识库。Dify 功能更全面，Agent 编排、工作流、API 发布能力更强，适合需要构建复杂 AI 应用的团队。简单知识库选 AnythingLLM，复杂应用选 Dify。",[25,1029,1030,1033],{},[41,1031,1032],{},"Q: 可以完全离线使用吗？","\nA: 可以。接 Ollama 本地模型 + 用本地 embedding 模型（如 bge-large-zh）+ 内置 LanceDB 向量库，整个系统完全离线运行，数据不出内网。适合数据敏感的企业场景。",[25,1035,1036,1039],{},[41,1037,1038],{},"Q: 免费开源版有什么限制？","\nA: MIT 协议开源版功能完整，无用户数 \u002F 文档数 \u002F API 调用限制。Cloud 版和 Enterprise 版主要是托管服务和企业管理功能（SSO \u002F 审计日志），功能层面开源版已够用。",[25,1041,1042,1045],{},[41,1043,1044],{},"Q: 支持中文文档吗？","\nA: 支持，但效果取决于 embedding 模型。默认 OpenAI embedding 对中文尚可，追求精度建议切换 bge-large-zh 或 m3e 模型。OCR 扫描件需额外配置 Tesseract 或接入外部 OCR 服务。",[20,1047,520],{"id":520},[25,1049,1050,527,1054,527,1056],{},[524,1051,1053],{"href":1052},"\u002Fagent\u002Fplatform\u002Fragflow.html","RAGFlow",[524,1055,531],{"href":530},[524,1057,1059],{"href":1058},"\u002Fagent\u002Fgeneral\u002Fperplexity.html","Perplexity",[20,1061,538],{"id":538},[151,1063,1064],{},[25,1065,543],{},[35,1067,1068,1074],{},[38,1069,1070],{},[524,1071,553],{"href":1072,"rel":1073},"https:\u002F\u002Fuseanything.com",[552],[38,1075,1076],{},[524,1077,560],{"href":1078,"rel":1079},"https:\u002F\u002Fgithub.com\u002FMintplex-Labs\u002Fanything-llm",[552],{"title":562,"searchDepth":563,"depth":563,"links":1081},[1082,1083,1084,1085,1086,1087,1088,1089,1090,1091,1092],{"id":22,"depth":566,"text":23},{"id":33,"depth":566,"text":33},{"id":95,"depth":566,"text":95},{"id":158,"depth":566,"text":159},{"id":221,"depth":566,"text":221},{"id":266,"depth":566,"text":266},{"id":405,"depth":566,"text":405},{"id":456,"depth":566,"text":457},{"id":492,"depth":566,"text":493},{"id":520,"depth":566,"text":520},{"id":538,"depth":566,"text":538},"\u002Fimg\u002Ftools\u002Fanythingllm.webp","AnythingLLM 真实评测：Mintplex Labs 出品的开源私有部署 LLM 平台（MIT 协议），一站式 RAG 知识库 + Agent + 多用户权限管理。支持 Docker\u002F桌面部署，接入 OpenAI\u002FClaude\u002FOllama 等任意模型，适合企业内网私有化 AI 知识库场景。",[583],{},"\u002Ftools\u002Fagent\u002Fplatform\u002Fanythingllm",[1099,1100,590,591],"windows","macos","Free \u002F 开源（MIT）\u002F Cloud",{"power":563,"ux":595,"price":596,"cn_support":563,"stability":563},{"title":616,"description":1094},"AnythingLLM - 开源私有部署 LLM 平台评测 | AIHO","agent\u002Fplatform\u002Fanythingllm",[1107,1108],{"title":553,"url":1072},{"title":560,"url":1078},"tools\u002Fagent\u002Fplatform\u002Fanythingllm","开源私有部署 LLM 平台，一站式 RAG + Agent + 多用户",[606,610,1112,1113,1114],"self-host","rag","multi-user","需要快速搭建私有化 AI 知识库且要求多用户权限管理的企业团队首选，MIT 协议 + 桌面\u002FDocker 双模式 + 任意模型接入降低了部署门槛，但 RAG 精度和 Agent 编排能力不及 Dify\u002FFastGPT 等专业平台。","iKAMhkQImqK_QZGWLaE4i_9IjAMMpCeSwojp4L34T6A",{"id":1118,"title":280,"alternatives":1119,"api_compatible":15,"body":1120,"category":577,"chinese_friendly":566,"cover":1570,"description":1571,"domestic":580,"extension":581,"faq":15,"free":580,"github":1555,"languages":1572,"lastVerified":584,"meta":1573,"models":15,"navigation":586,"notSuitable":15,"opensource":586,"path":1574,"pillar":588,"platforms":1575,"priceTable":15,"pricing":1576,"published":593,"relatedPlaybooks":15,"relatedReviews":15,"score":1577,"self_host":580,"seo":1578,"seoTitle":1579,"slug":13,"sources":1580,"stem":1583,"suitable":15,"tagline":1584,"tags":1585,"updated":584,"verdict":1587,"website":1549,"__hash__":1588},"tools\u002Ftools\u002Fagent\u002Fplatform\u002Fautogen.md",[599,12,618],{"type":17,"value":1121,"toc":1557},[1122,1124,1131,1134,1136,1190,1192,1195,1197,1203,1227,1231,1254,1256,1289,1291,1403,1405,1453,1455,1486,1488,1494,1512,1518,1526,1528,1537,1539,1543],[20,1123,23],{"id":22},[25,1125,1126,1127,1130],{},"AutoGen 是微软研究院开源的多 Agent 对话框架（MIT 协议），用 Python 代码定义 Agent 角色、对话流程和工具调用。核心是 ",[228,1128,1129],{},"ConversableAgent"," + Group Chat 模式——多个 Agent 自动对话协作完成任务，支持代码执行、工具调用、人在回路（human-in-the-loop）。AutoGen 0.4+ 重构为事件驱动架构，性能和扩展性大幅提升。",[25,1132,1133],{},"适合：AI 研究者、需要精细控制 Agent 协作逻辑的高级开发者、多 Agent 实验项目。不适合：快速原型验证（用 CrewAI）、非技术用户（用 Dify \u002F Flowise）、需要 GUI 的场景、追求 API 稳定性的生产项目。",[20,1135,33],{"id":33},[35,1137,1138,1144,1149,1155,1161,1167,1173,1178,1184],{},[38,1139,1140,1143],{},[41,1141,1142],{},"多 Agent 对话","：Group Chat 模式，多个 Agent 自动对话协作完成任务",[38,1145,1146,1148],{},[41,1147,343],{},"：内置 Docker 代码执行器，Agent 可写代码 + 运行 + 调试",[38,1150,1151,1154],{},[41,1152,1153],{},"工具调用","：自定义函数工具，Agent 自动选择和调用",[38,1156,1157,1160],{},[41,1158,1159],{},"人在回路","：Human-in-the-loop 模式，关键决策需人工确认",[38,1162,1163,1166],{},[41,1164,1165],{},"事件驱动架构","：0.4+ 重构为 async 事件驱动，支持分布式 Agent",[38,1168,1169,1172],{},[41,1170,1171],{},"可定制 Agent","：system message \u002F 工具集 \u002F 终止条件全可自定义",[38,1174,1175,1177],{},[41,1176,79],{},"：OpenAI \u002F Claude \u002F Azure \u002F Ollama \u002F Gemini 等",[38,1179,1180,1183],{},[41,1181,1182],{},"Agent 可组合","：嵌套 Agent、层级 Agent、条件路由",[38,1185,1186,1189],{},[41,1187,1188],{},"可观测性","：集成 OpenTelemetry \u002F LangSmith 追踪 Agent 行为",[20,1191,95],{"id":95},[25,1193,1194],{},"完全免费、MIT 开源、商用免费。运行成本仅来自所接入的 LLM API 调用费用。",[20,1196,159],{"id":158},[151,1198,1199],{},[25,1200,164,1201],{},[41,1202,167],{},[35,1204,1205,1208,1211,1214,1221,1224],{},[38,1206,1207],{},"Group Chat 模式让多 Agent 协作真正\"自动化\"，代码 reviewer + coder + tester 角色分工清晰",[38,1209,1210],{},"Docker 代码执行器安全隔离，Agent 写的代码在沙箱中运行",[38,1212,1213],{},"0.4+ 的事件驱动架构性能提升明显，异步并发能力强",[38,1215,1216,1217,1220],{},"自定义工具集成灵活，Python 函数加 ",[228,1218,1219],{},"@user_function"," 装饰器即可",[38,1222,1223],{},"微软背书，学术认可度高，论文引用多",[38,1225,1226],{},"人在回路模式适合需要人工把关的高风险场景",[25,1228,1229],{},[41,1230,195],{},[35,1232,1233,1236,1239,1242,1245,1248,1251],{},[38,1234,1235],{},"学习曲线非常陡峭，文档虽全但概念密集，新手容易劝退",[38,1237,1238],{},"0.2 → 0.4 API 大改，迁移成本高，网上旧教程大量失效",[38,1240,1241],{},"Agent 对话容易\"跑飞\"——无限循环 \u002F 偏离主题，需仔细设计终止条件",[38,1243,1244],{},"无 GUI，调试全靠日志和 print，排查多 Agent 对话链路费时",[38,1246,1247],{},"Token 消耗大——多 Agent 对话轮次多，API 费用叠加明显",[38,1249,1250],{},"错误处理不够健壮，LLM 返回格式异常时容易崩溃",[38,1252,1253],{},"社区活跃度不如 LangChain \u002F CrewAI，遇到问题搜索不到答案",[20,1255,221],{"id":221},[223,1257,1258,1264,1269,1274,1280,1286],{},[38,1259,1260,1263],{},[228,1261,1262],{},"pip install autogen-agentchat autogen-ext","（0.4+ 新包名）",[38,1265,234,1266,1268],{},[228,1267,237],{}," 环境变量或代码内传入",[38,1270,241,1271],{},[228,1272,1273],{},"AssistantAgent(name=\"coder\", system_message=\"...\", model_client=...)",[38,1275,1276,1277],{},"创建 Group Chat：",[228,1278,1279],{},"RoundRobinGroupChat(agents=[agent1, agent2])",[38,1281,1282,1283],{},"发起任务：",[228,1284,1285],{},"result = await team.run(task=\"写一个贪吃蛇游戏\")",[38,1287,1288],{},"进阶：加 Docker 代码执行器 + 自定义工具 + 人在回路",[20,1290,266],{"id":266},[97,1292,1293,1307],{},[100,1294,1295],{},[103,1296,1297,1299,1301,1303,1305],{},[106,1298,275],{},[106,1300,280],{},[106,1302,10],{},[106,1304,283],{},[106,1306,286],{},[115,1308,1309,1324,1336,1348,1360,1373,1389],{},[103,1310,1311,1314,1317,1319,1321],{},[120,1312,1313],{},"形态",[120,1315,1316],{},"Python 框架",[120,1318,1316],{},[120,1320,1316],{},[120,1322,1323],{},"可视化平台",[103,1325,1326,1328,1330,1332,1334],{},[120,1327,293],{},[120,1329,299],{},[120,1331,855],{},[120,1333,299],{},[120,1335,296],{},[103,1337,1338,1340,1342,1344,1346],{},[120,1339,326],{},[120,1341,332],{},[120,1343,329],{},[120,1345,335],{},[120,1347,338],{},[103,1349,1350,1352,1354,1356,1358],{},[120,1351,343],{},[120,1353,349],{},[120,1355,346],{},[120,1357,346],{},[120,1359,354],{},[103,1361,1362,1364,1366,1369,1371],{},[120,1363,375],{},[120,1365,381],{},[120,1367,1368],{},"❌（有 CrewAI Studio）",[120,1370,381],{},[120,1372,370],{},[103,1374,1375,1378,1381,1384,1387],{},[120,1376,1377],{},"API 稳定性",[120,1379,1380],{},"一般（大改过）",[120,1382,1383],{},"较好",[120,1385,1386],{},"好",[120,1388,1386],{},[103,1390,1391,1393,1396,1398,1401],{},[120,1392,390],{},[120,1394,1395],{},"研究 \u002F 复杂协作",[120,1397,393],{},[120,1399,1400],{},"精确流程控制",[120,1402,402],{},[20,1404,405],{"id":405},[35,1406,1407,1417,1423,1429,1435,1441,1447],{},[38,1408,1409,1412,1413,1416],{},[41,1410,1411],{},"锁定版本","：0.2 和 0.4 API 不兼容，",[228,1414,1415],{},"pip install"," 时务必指定版本",[38,1418,1419,1422],{},[41,1420,1421],{},"设计终止条件","：Group Chat 不设终止条件会无限对话，设 max_turns + 终止关键词",[38,1424,1425,1428],{},[41,1426,1427],{},"Token 成本控制","：多 Agent 对话 token 消耗是单 Agent 的 3-5 倍，用 GPT-4 级模型注意费用",[38,1430,1431,1434],{},[41,1432,1433],{},"代码执行器一定要用 Docker","：直接本地执行 Agent 生成的代码有安全风险",[38,1436,1437,1440],{},[41,1438,1439],{},"别指望第一次跑通","：system message 调试 + 工具定义 + 终止条件需要反复迭代",[38,1442,1443,1446],{},[41,1444,1445],{},"错误处理要完善","：LLM 返回异常格式时手动 catch + 重试",[38,1448,1449,1452],{},[41,1450,1451],{},"不要用旧教程","：0.4+ 完全重构，网上大部分 AutoGen 教程是 0.2 版本的",[20,1454,457],{"id":456},[35,1456,1457,1460,1463,1466,1469,1472,1475,1477,1480,1483],{},[38,1458,1459],{},"✅ AI 研究者实验多 Agent 协作模式",[38,1461,1462],{},"✅ 需要代码级精细控制 Agent 行为",[38,1464,1465],{},"✅ 需要 Agent 代码执行 + 自动调试",[38,1467,1468],{},"✅ 人在回路的高风险决策场景",[38,1470,1471],{},"✅ 学术项目 \u002F 论文复现",[38,1473,1474],{},"❌ 快速原型验证（用 CrewAI，API 更简洁）",[38,1476,477],{},[38,1478,1479],{},"❌ 追求 API 稳定性的生产项目（版本变动大）",[38,1481,1482],{},"❌ 需要可视化调试（无 GUI，全靠日志）",[38,1484,1485],{},"❌ 预算敏感场景（多 Agent 对话 token 消耗大）",[20,1487,493],{"id":492},[25,1489,1490,1493],{},[41,1491,1492],{},"Q: AutoGen 和 CrewAI 怎么选？","\nA: AutoGen 更底层、更灵活，适合研究和复杂多 Agent 协作实验，但学习成本高。CrewAI API 更简洁直观，角色 + 任务 + 流程的概念更易理解，适合业务自动化场景。研究选 AutoGen，做产品选 CrewAI。",[25,1495,1496,1499,1500,1503,1504,1507,1508,1511],{},[41,1497,1498],{},"Q: AutoGen 0.2 和 0.4 有什么区别？","\nA: 0.4 是完全重构版本——从同步改为异步事件驱动架构，包名从 ",[228,1501,1502],{},"pyautogen"," 改为 ",[228,1505,1506],{},"autogen-agentchat"," + ",[228,1509,1510],{},"autogen-ext","，API 全面更新。性能和扩展性大幅提升但旧代码无法直接迁移。新项目直接用 0.4+。",[25,1513,1514,1517],{},[41,1515,1516],{},"Q: 多 Agent 对话成本高吗？","\nA: 高。多 Agent 每轮对话都消耗 token，一个任务 5-10 轮对话是常态，使用 GPT-4 级模型单个任务可能花费 $0.5-2。建议开发调试用便宜模型（GPT-4o-mini），生产再切高级模型。",[25,1519,1520,1522,1523,1525],{},[41,1521,510],{},"\nA: 可以。通过 ",[228,1524,1510],{}," 的 OpenAI 兼容客户端接入 Ollama \u002F vLLM \u002F LM Studio 的本地模型端点。但本地模型能力有限，复杂多 Agent 协作效果可能不如 GPT-4 \u002F Claude。",[20,1527,520],{"id":520},[25,1529,1530,527,1533,527,1535],{},[524,1531,10],{"href":1532},"\u002Fagent\u002Fplatform\u002Fcrewai.html",[524,1534,531],{"href":530},[524,1536,535],{"href":534},[20,1538,538],{"id":538},[151,1540,1541],{},[25,1542,543],{},[35,1544,1545,1551],{},[38,1546,1547],{},[524,1548,553],{"href":1549,"rel":1550},"https:\u002F\u002Fmicrosoft.github.io\u002Fautogen",[552],[38,1552,1553],{},[524,1554,560],{"href":1555,"rel":1556},"https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fautogen",[552],{"title":562,"searchDepth":563,"depth":563,"links":1558},[1559,1560,1561,1562,1563,1564,1565,1566,1567,1568,1569],{"id":22,"depth":566,"text":23},{"id":33,"depth":566,"text":33},{"id":95,"depth":566,"text":95},{"id":158,"depth":566,"text":159},{"id":221,"depth":566,"text":221},{"id":266,"depth":566,"text":266},{"id":405,"depth":566,"text":405},{"id":456,"depth":566,"text":457},{"id":492,"depth":566,"text":493},{"id":520,"depth":566,"text":520},{"id":538,"depth":566,"text":538},"\u002Fimg\u002Ftools\u002Fautogen.webp","AutoGen 真实评测：微软开源的多 Agent 对话框架（MIT 协议），通过代码定义 Agent 角色和协作流程，支持多 Agent 对话、工具调用、代码执行。适合需要精细控制多 Agent 协作逻辑的开发者和研究团队。",[583],{},"\u002Ftools\u002Fagent\u002Fplatform\u002Fautogen",[590,591],"Free \u002F 开源（MIT）",{"power":595,"ux":566,"price":596,"cn_support":566,"stability":563},{"title":280,"description":1571},"AutoGen - 微软多 Agent 框架评测与使用 | AIHO",[1581,1582],{"title":553,"url":1549},{"title":560,"url":1555},"tools\u002Fagent\u002Fplatform\u002Fautogen","微软开源多 Agent 对话框架，代码驱动 Agent 协作",[606,607,608,1586,609,610],"microsoft","需要代码级精细控制多 Agent 协作逻辑的研究者和高级开发者首选，微软背书 + 代码执行 + Group Chat 模式强大，但学习曲线陡峭、API 稳定性一般、无 GUI，不适合快速原型或非技术用户。","Asmg5F2Nn2KaD4XJWzqRwFA4s1iK_qUmi34M41yhcNE",{"id":1590,"title":1591,"alternatives":1592,"api_compatible":1594,"body":1595,"category":577,"chinese_friendly":596,"cover":2423,"description":2424,"domestic":580,"extension":581,"faq":15,"free":580,"github":15,"languages":2425,"lastVerified":15,"meta":2427,"models":2428,"navigation":586,"notSuitable":2434,"opensource":580,"path":2438,"pillar":588,"platforms":2439,"priceTable":2441,"pricing":2462,"published":2463,"relatedPlaybooks":15,"relatedReviews":2464,"score":2469,"self_host":580,"seo":2470,"seoTitle":2471,"slug":2472,"sources":2473,"stem":2484,"suitable":2485,"tagline":2491,"tags":2492,"updated":2498,"verdict":2499,"website":1617,"__hash__":2500},"tools\u002Ftools\u002Fagent\u002Fplatform\u002Fcoze.md","Coze",[618,619,1593],"agent\u002Fplatform\u002Fyuanqi",[],{"type":17,"value":1596,"toc":2407},[1597,1599,1640,1645,1648,1653,1656,1670,1679,1729,1738,1742,1745,1776,1784,1788,1791,1811,1815,1895,1905,1908,1911,1931,1940,1944,1980,1983,1986,2126,2142,2173,2176,2249,2251,2254,2271,2274,2303,2305,2370,2372,2399],[20,1598,23],{"id":22},[1600,1601,1606,1627],"div",{"className":1602},[1603,1604,1605],"card","p-5","my-4",[25,1607,1608,1611,1612,1626],{},[41,1609,1610],{},"一句话："," 字节跳动出品的低代码 Agent 平台，",[41,1613,1614,1615,1620,1621],{},"国内版 ",[524,1616,1619],{"href":1617,"rel":1618},"https:\u002F\u002Fwww.coze.cn",[552],"coze.cn","（中文叫\"扣子\"）+ 国际版 ",[524,1622,1625],{"href":1623,"rel":1624},"https:\u002F\u002Fwww.coze.com",[552],"coze.com"," 双轨运营。国内版深度集成飞书 \u002F 抖音生态、原生接入豆包；国际版集成 OpenAI \u002F Claude \u002F Gemini。",[25,1628,1629,1630,1633,1634,1639],{},"最大价值是 ",[41,1631,1632],{},"零代码、最快上手","——文档详细 + 中文社区活跃 + 模板多，根据 ",[524,1635,1638],{"href":1636,"rel":1637},"https:\u002F\u002Fwww.cnblogs.com\u002Fuulucias\u002Fp\u002F19449008",[552],"博客园 2026-01 选型指南"," 引用的真实案例，\"某电商公司用 Coze 搭建客服机器人，3 天上线，月成本 \u003C 1000 元\"。",[151,1641,1642],{},[25,1643,1644],{},"来源说明：本文基于 coze.cn \u002F coze.com 官方页面、docs.coze.com 文档、第三方选型评测（cnblogs \u002F besthub \u002F aibotgo）综合整理。字节产品迭代很快，价格 \u002F 功能请以最新官方页面为准。",[20,1646,1647],{"id":1647},"核心特性",[1649,1650,1652],"h3",{"id":1651},"可视化工作流最大卖点","可视化工作流（最大卖点）",[25,1654,1655],{},"Coze 把 Agent 拆成两层：",[35,1657,1658,1664],{},[38,1659,1660,1663],{},[41,1661,1662],{},"Bot \u002F 智能体","：对话式 AI，配置 prompt + 知识库 + 插件",[38,1665,1666,1669],{},[41,1667,1668],{},"工作流（Workflow）","：DAG 节点编排，可被 Bot 调用，也可独立部署",[25,1671,1672,1673,1678],{},"工作流节点类型（基于 ",[524,1674,1677],{"href":1675,"rel":1676},"https:\u002F\u002Fdeveloper.volcengine.com\u002Farticles\u002F7530117616687480851",[552],"火山引擎社区 2025 实战","）：",[35,1680,1681,1687,1693,1699,1705,1711,1717,1723],{},[38,1682,1683,1686],{},[41,1684,1685],{},"开始 \u002F 结束节点","：输入输出",[38,1688,1689,1692],{},[41,1690,1691],{},"大模型节点","：调豆包 \u002F GPT \u002F Claude 任一模型",[38,1694,1695,1698],{},[41,1696,1697],{},"代码节点","：内嵌 Python \u002F JavaScript（飞书插件常需要数据格式转换）",[38,1700,1701,1704],{},[41,1702,1703],{},"循环节点","：批量处理多条数据",[38,1706,1707,1710],{},[41,1708,1709],{},"条件节点","：分支判断",[38,1712,1713,1716],{},[41,1714,1715],{},"插件节点","：调 Coze 插件市场的工具",[38,1718,1719,1722],{},[41,1720,1721],{},"知识库节点","：RAG 检索",[38,1724,1725,1728],{},[41,1726,1727],{},"HTTP 节点","：调外部 API",[25,1730,1731,1732,1737],{},"典型用例（参考 ",[524,1733,1736],{"href":1734,"rel":1735},"https:\u002F\u002Fwww.toutiao.com\u002Farticle\u002F7469986334686315017",[552],"今日头条 涛哥讲AI 2025-02 教程","）：读飞书多维表格 → 批量调大模型转写为小红书风格 → 写回飞书。整个流程零代码完成。",[1649,1739,1741],{"id":1740},"插件市场200-官方-海量第三方","插件市场（200+ 官方 + 海量第三方）",[25,1743,1744],{},"Coze 的\"插件\"是封装好的 API 工具，比如：",[35,1746,1747,1753,1759,1765,1770],{},[38,1748,1749,1752],{},[41,1750,1751],{},"飞书多维表格","：增删改查记录（国内 toB 场景的杀手锏）",[38,1754,1755,1758],{},[41,1756,1757],{},"图像生成","：调豆包 \u002F SD \u002F DALL-E",[38,1760,1761,1764],{},[41,1762,1763],{},"联网搜索","：实时网页检索",[38,1766,1767,1769],{},[41,1768,343],{},"：在线运行 Python",[38,1771,1772,1775],{},[41,1773,1774],{},"第三方 SaaS","：微博、抖音、bilibili、Notion……",[25,1777,1778,1783],{},[524,1779,1782],{"href":1780,"rel":1781},"https:\u002F\u002Fkouziai.github.io\u002F",[552],"扣子空间介绍"," 提到：\"插件数量突破 500 个\"——可信度待官方确认，但量级在百级别是确定的。",[1649,1785,1787],{"id":1786},"bot-商店-多平台一键发布","Bot 商店 + 多平台一键发布",[25,1789,1790],{},"发布渠道：",[35,1792,1793,1796,1799,1802,1805,1808],{},[38,1794,1795],{},"飞书机器人（一键绑）",[38,1797,1798],{},"抖音 \u002F 头条号",[38,1800,1801],{},"微信小程序 \u002F 公众号（部分需企业认证）",[38,1803,1804],{},"自定义网页嵌入",[38,1806,1807],{},"API 接口（提供 OpenAPI 风格 REST 调用）",[38,1809,1810],{},"Discord（国际版）",[1649,1812,1814],{"id":1813},"国内版-vs-国际版","国内版 vs 国际版",[97,1816,1817,1829],{},[100,1818,1819],{},[103,1820,1821,1823,1826],{},[106,1822,275],{},[106,1824,1825],{},"扣子（coze.cn）",[106,1827,1828],{},"Coze（coze.com）",[115,1830,1831,1842,1852,1862,1873,1884],{},[103,1832,1833,1836,1839],{},[120,1834,1835],{},"主力模型",[120,1837,1838],{},"豆包 Pro \u002F DeepSeek \u002F Qwen \u002F Kimi",[120,1840,1841],{},"OpenAI \u002F Claude \u002F Gemini \u002F Cohere",[103,1843,1844,1847,1850],{},[120,1845,1846],{},"飞书 \u002F 抖音 \u002F 微信集成",[120,1848,1849],{},"✅ 原生",[120,1851,381],{},[103,1853,1854,1857,1860],{},[120,1855,1856],{},"Discord \u002F Slack 集成",[120,1858,1859],{},"⚠️ 有限",[120,1861,1849],{},[103,1863,1864,1867,1870],{},[120,1865,1866],{},"数据存储位置",[120,1868,1869],{},"国内",[120,1871,1872],{},"海外",[103,1874,1875,1878,1881],{},[120,1876,1877],{},"支付",[120,1879,1880],{},"微信 \u002F 支付宝",[120,1882,1883],{},"海外信用卡",[103,1885,1886,1889,1892],{},[120,1887,1888],{},"内容合规",[120,1890,1891],{},"严格审核",[120,1893,1894],{},"宽松",[25,1896,1897,1900,1901,1904],{},[41,1898,1899],{},"实践建议","：国内 toC \u002F toB 用扣子，海外项目 \u002F 接 GPT 用 coze.com。两边账号 \u002F 工作流 ",[41,1902,1903],{},"不互通","。",[20,1906,1907],{"id":1907},"价格与运行成本",[25,1909,1910],{},"国内版（扣子）：",[35,1912,1913,1919,1925],{},[38,1914,1915,1918],{},[41,1916,1917],{},"免费版","：免费模型有日额度（豆包 lite 等），适合个人玩 \u002F Demo",[38,1920,1921,1924],{},[41,1922,1923],{},"专业版","：按调用计费，模型 + 高并发，单 token 价比直连 API 略贵但省事",[38,1926,1927,1930],{},[41,1928,1929],{},"企业版","：议价，含 VPC、私有化（限定场景）、SLA",[25,1932,1933,1934,1939],{},"国际版（coze.com）的定价模式据 ",[524,1935,1938],{"href":1936,"rel":1937},"https:\u002F\u002Fdocs.coze.com\u002F",[552],"官方文档"," 描述：\"按你访问和使用的功能分别计费，每个功能有自己的计费模型\"——目前没有简单的\"$X\u002F月\"档位，类似按 token \u002F 工具调用的 metered billing。",[20,1941,1943],{"id":1942},"上手-10-分钟","上手 10 分钟",[223,1945,1946,1959,1962,1965,1968,1971,1974,1977],{},[38,1947,1948,1949,1953,1954,1958],{},"打开 ",[524,1950,1952],{"href":1617,"rel":1951},[552],"www.coze.cn","（国内）或 ",[524,1955,1957],{"href":1623,"rel":1956},[552],"www.coze.com","（国际），用飞书 \u002F 抖音账号 \u002F Google 账号登录",[38,1960,1961],{},"左侧\"工作空间\" → \"+创建 Bot\"，起个名字",[38,1963,1964],{},"选模型（国内推荐豆包 Pro，国际推 Claude Sonnet 4）",[38,1966,1967],{},"写 Bot 角色 prompt",[38,1969,1970],{},"可选：上传 PDF \u002F 文档建知识库",[38,1972,1973],{},"测试一下对话效果",[38,1975,1976],{},"右上\"发布\" → 选渠道（飞书 \u002F 抖音 \u002F API \u002F Web）",[38,1978,1979],{},"拿到调用 URL \u002F 飞书机器人 webhook",[25,1981,1982],{},"进阶：在\"资源库\"创建工作流，拖节点 → 调试 → 在 Bot 里\"添加工作流\"引用。",[20,1984,1985],{"id":1985},"与同类怎么选",[97,1987,1988,2012],{},[100,1989,1990],{},[103,1991,1992,1994,1996,2001,2006],{},[106,1993,275],{},[106,1995,1591],{},[106,1997,1998],{},[524,1999,286],{"href":2000},"\u002Fagent\u002Fplatform\u002Fdify.html",[106,2002,2003],{},[524,2004,840],{"href":2005},"\u002Fagent\u002Fplatform\u002Ffastgpt.html",[106,2007,2008],{},[524,2009,2011],{"href":2010},"\u002Fagent\u002Fplatform\u002Fyuanqi.html","元器 yuanqi",[115,2013,2014,2027,2042,2057,2073,2085,2099,2113],{},[103,2015,2016,2019,2021,2023,2025],{},[120,2017,2018],{},"开源",[120,2020,381],{},[120,2022,370],{},[120,2024,370],{},[120,2026,381],{},[103,2028,2029,2032,2035,2037,2039],{},[120,2030,2031],{},"私有部署",[120,2033,2034],{},"⚠️ 仅企业版",[120,2036,370],{},[120,2038,370],{},[120,2040,2041],{},"⚠️",[103,2043,2044,2046,2049,2052,2054],{},[120,2045,293],{},[120,2047,2048],{},"★ 最简单",[120,2050,2051],{},"★★★",[120,2053,2051],{},[120,2055,2056],{},"★★",[103,2058,2059,2062,2065,2068,2071],{},[120,2060,2061],{},"工作流编排",[120,2063,2064],{},"★★★★☆",[120,2066,2067],{},"★★★★★",[120,2069,2070],{},"★★★☆☆",[120,2072,2070],{},[103,2074,2075,2077,2079,2081,2083],{},[120,2076,878],{},[120,2078,2070],{},[120,2080,2064],{},[120,2082,2067],{},[120,2084,2070],{},[103,2086,2087,2090,2092,2094,2096],{},[120,2088,2089],{},"字节生态",[120,2091,2067],{},[120,2093,381],{},[120,2095,381],{},[120,2097,2098],{},"❌（腾讯系）",[103,2100,2101,2104,2107,2109,2111],{},[120,2102,2103],{},"插件市场",[120,2105,2106],{},"★★★★★ 200+",[120,2108,2070],{},[120,2110,2070],{},[120,2112,2070],{},[103,2114,2115,2118,2120,2122,2124],{},[120,2116,2117],{},"中文社区",[120,2119,2067],{},[120,2121,2064],{},[120,2123,2064],{},[120,2125,2064],{},[25,2127,2128,2131,2132,2137,2138,1678],{},[41,2129,2130],{},"怎么选","（综合 ",[524,2133,2136],{"href":2134,"rel":2135},"https:\u002F\u002Fwww.besthub.dev\u002Farticles\u002Fcoze-vs-dify-vs-fastgpt-which-ai-agent-platform-fits-your-needs-fa59cf97b798",[552],"BestHub 2025-07"," 和 ",[524,2139,2141],{"href":1636,"rel":2140},[552],"博客园 2026-01",[35,2143,2144,2150,2158,2165],{},[38,2145,2146,2149],{},[41,2147,2148],{},"快速验证 \u002F 不懂代码 \u002F 1-2 天出原型"," → Coze",[38,2151,2152,2155,2156],{},[41,2153,2154],{},"数据安全要求高 \u002F 复杂业务流程 \u002F 有技术团队"," → ",[524,2157,286],{"href":2000},[38,2159,2160,2155,2163],{},[41,2161,2162],{},"核心场景就是企业知识库 QA",[524,2164,840],{"href":2005},[38,2166,2167,2155,2170],{},[41,2168,2169],{},"QQ \u002F 微信生态 + 腾讯系",[524,2171,2172],{"href":2010},"元器",[20,2174,2175],{"id":2175},"避坑清单",[35,2177,2178,2188,2194,2209,2219,2225,2237,2243],{},[38,2179,2180,2183,2184,2187],{},[41,2181,2182],{},"国内版 vs 国际版的\"双账号陷阱\"","：扣子（coze.cn）和 Coze（coze.com）是",[41,2185,2186],{},"两套独立系统","，账号、Bot、工作流不互通；想\"国内调通后搬到海外\"需要重新搭",[38,2189,2190,2193],{},[41,2191,2192],{},"专业版按调用计费容易超预算","：上线前一定在测试环境跑量估算月成本，否则爆款 Bot 一夜烧爆账户",[38,2195,2196,2199,2200,2203,2204,2208],{},[41,2197,2198],{},"飞书多维表格插件数据格式坑","：写入多维表格需要 ",[228,2201,2202],{},"Array\u003CObject>"," 格式，代码节点要做转换（参考 ",[524,2205,2207],{"href":1675,"rel":2206},[552],"火山引擎 2025 教程","）",[38,2210,2211,2214,2215,2218],{},[41,2212,2213],{},"工作流读取飞书表格默认 20 条","：要改 ",[228,2216,2217],{},"page_size","，最大 500 条；超过 500 要分页或循环",[38,2220,2221,2224],{},[41,2222,2223],{},"运行超时","：单工作流执行有时间上限，记录条数 > 50 时建议在 Bot 里\"异步\"调用，不要直接走工作流",[38,2226,2227,2230,2231,2233,2234,2236],{},[41,2228,2229],{},"企业版\"私有化\"是有限的","：完全数据不出网仍建议 ",[524,2232,286],{"href":2000}," \u002F ",[524,2235,840],{"href":2005}," 自托管",[38,2238,2239,2242],{},[41,2240,2241],{},"国际版接 Claude \u002F GPT 需要 BYOK","：自己绑海外信用卡，平台不代付",[38,2244,2245,2248],{},[41,2246,2247],{},"审核合规","：国内版对 prompt \u002F 输出有内容审核，金融 \u002F 医疗 \u002F 政治话题可能被拦",[20,2250,457],{"id":456},[25,2252,2253],{},"✅ 适合：",[35,2255,2256,2259,2262,2265,2268],{},[38,2257,2258],{},"产品 \u002F 运营 \u002F 非技术人员快速做 Bot",[38,2260,2261],{},"在飞书 \u002F 抖音 \u002F 头条生态内做集成",[38,2263,2264],{},"个人副业（小红书账号批量内容生成等）",[38,2266,2267],{},"中小企业客服 Bot（3 天上线）",[38,2269,2270],{},"想用豆包 \u002F DeepSeek 国产模型的人",[25,2272,2273],{},"❌ 不适合：",[35,2275,2276,2279,2282,2287,2294],{},[38,2277,2278],{},"金融 \u002F 政府 \u002F 医疗（数据敏感，需自托管）",[38,2280,2281],{},"复杂业务系统深度集成（自由度不够）",[38,2283,2284,2285,2208],{},"反感字节生态（去 ",[524,2286,286],{"href":2000},[38,2288,2289,2290,2233,2292,2208],{},"希望开源 \u002F 完全自主可控（去 ",[524,2291,286],{"href":2000},[524,2293,840],{"href":2005},[38,2295,2296,2297,2233,2299,2302],{},"海外 toB SaaS 产品后端（",[524,2298,286],{"href":2000},[524,2300,843],{"href":2301},"\u002Fagent\u002Fplatform\u002Flangflow.html"," 更合适）",[20,2304,520],{"id":520},[35,2306,2307,2322,2340,2359],{},[38,2308,2309,2310,2233,2312,2233,2314,2233,2316,2233,2318],{},"同类对比：",[524,2311,286],{"href":2000},[524,2313,840],{"href":2005},[524,2315,2172],{"href":2010},[524,2317,843],{"href":2301},[524,2319,2321],{"href":2320},"\u002Fagent\u002Fplatform\u002Fn8n.html","n8n",[38,2323,2324,2325,2233,2329,2233,2333,2233,2337],{},"概念：",[524,2326,2328],{"href":2327},"\u002Fwiki\u002Fai-agent.html","AI Agent",[524,2330,2332],{"href":2331},"\u002Fwiki\u002Frag.html","RAG",[524,2334,2336],{"href":2335},"\u002Fwiki\u002Ffunction-calling.html","Function Calling",[524,2338,2339],{"href":2327},"Multi-Agent",[38,2341,2342,2343,2233,2347,2233,2351,2233,2355],{},"模型：",[524,2344,2346],{"href":2345},"\u002Fmodels\u002Fdoubao-1-5-pro.html","豆包 Doubao",[524,2348,2350],{"href":2349},"\u002Fmodels\u002Fdeepseek-v3.html","DeepSeek-V3",[524,2352,2354],{"href":2353},"\u002Fmodels\u002Fqwen-3.html","Qwen3",[524,2356,2358],{"href":2357},"\u002Fmodels\u002Fglm-5.2.html","GLM-5.2",[38,2360,2361,2362,2233,2366],{},"进阶：",[524,2363,2365],{"href":2364},"\u002Fwiki\u002Fprompt-engineering.html","Prompt Engineering",[524,2367,2369],{"href":2368},"\u002Fwiki\u002Fcontext-engineering.html","Context Engineering",[20,2371,538],{"id":538},[35,2373,2374,2380,2386,2393,2396],{},[38,2375,2376,2377],{},"国内版：",[524,2378,1617],{"href":1617,"rel":2379},[552],[38,2381,2382,2383],{},"国际版：",[524,2384,1623],{"href":1623,"rel":2385},[552],[38,2387,2388,2389],{},"官方文档：",[524,2390,2391],{"href":2391,"rel":2392},"https:\u002F\u002Fdocs.coze.com",[552],[38,2394,2395],{},"第三方选型评测：cnblogs.com \u002F besthub.dev \u002F aibotgo.net",[38,2397,2398],{},"实战教程：火山引擎社区、今日头条 涛哥讲AI",[25,2400,2401,2402,2406],{},"本卡片由 AIHO 编辑部根据官方公开资料与第三方评测整理。所有事实点均标注来源；如发现价格 \u002F 功能 \u002F 渠道与最新官方信息不一致，请通过 ",[524,2403,2405],{"href":2404},"mailto:hello@aiho.net","反馈邮箱"," 反馈。",{"title":562,"searchDepth":563,"depth":563,"links":2408},[2409,2410,2416,2417,2418,2419,2420,2421,2422],{"id":22,"depth":566,"text":23},{"id":1647,"depth":566,"text":1647,"children":2411},[2412,2413,2414,2415],{"id":1651,"depth":563,"text":1652},{"id":1740,"depth":563,"text":1741},{"id":1786,"depth":563,"text":1787},{"id":1813,"depth":563,"text":1814},{"id":1907,"depth":566,"text":1907},{"id":1942,"depth":566,"text":1943},{"id":1985,"depth":566,"text":1985},{"id":2175,"depth":566,"text":2175},{"id":456,"depth":566,"text":457},{"id":520,"depth":566,"text":520},{"id":538,"depth":566,"text":538},"\u002Fimg\u002Ftools\u002Fcoze.webp","Coze 扣子 2026 真实评测：字节低代码 AI Agent 平台，支持可视化工作流、插件市场、知识库和 Bot 发布。本文对比 Dify、FastGPT、n8n，梳理国内版\u002F国际版差异、价格、适合场景和避坑建议。",[2426,583],"zh",{},[2429,2430,2431,2432,2433],"doubao-pro","deepseek-v3","qwen-max","gpt-4o (国际版)","claude (国际版)",[2435,2436,2437],"需要私有部署、数据不出网（去 Dify \u002F FastGPT）","需要深度定制（去 LangGraph \u002F n8n）","对字节生态依赖反感","\u002Ftools\u002Fagent\u002Fplatform\u002Fcoze",[2440],"web",[2442,2448,2453,2457],{"plan":2443,"price":2444,"limit":2445,"cn_pay":2446,"note":2447},"免费版（扣子）","¥0","免费模型限额 + 基础功能","—","个人试水 \u002F 1-2 天 MVP",{"plan":1923,"price":2449,"limit":2450,"cn_pay":2451,"note":2452},"按调用计费","高级模型 + 高并发 + 商用授权","✅ 微信\u002F支付宝","上量后切",{"plan":1929,"price":2454,"limit":2455,"cn_pay":370,"note":2456},"议价","VPC、合规、SLA、私有化（部分）","B 端落地",{"plan":2458,"price":2459,"cn_pay":2460,"note":2461,"limit":2446},"国际版 (coze.com)","免费起步 + 用量计费","需海外卡","可接 GPT\u002FClaude\u002FGemini","免费档 \u002F 专业版按调用计费 \u002F 企业版议价","2026-06-18",[2465,2466,2467,2468],"coze-deep-review","coze-vs-dify","dify-deep-review","fastgpt-deep-review",{"power":595,"ux":596,"price":595,"cn_support":596,"stability":595},{"title":1591,"description":2424},"Coze 扣子评测 2026：字节 AI Agent 平台，对比 Dify","agent\u002Fplatform\u002Fcoze",[2474,2476,2478,2480,2482],{"title":2475,"url":1617},"Coze 国内版（扣子）",{"title":2477,"url":1623},"Coze 国际版",{"title":2479,"url":2391},"Coze 官方文档",{"title":2481,"url":1636},"Coze vs Dify vs FastGPT 选型 2026",{"title":2483,"url":2134},"BestHub 三平台对比","tools\u002Fagent\u002Fplatform\u002Fcoze",[2486,2487,2488,2489,2490],"想 1 小时做出一个 Bot 的产品 \u002F 运营","需要发布到飞书 \u002F 微信 \u002F 抖音的 Bot","工作流可视化编排（不想写代码）","需要批量调用国内大模型 + 飞书多维表格的工作流","C 端 \u002F 轻量 toB 场景","字节出品的 Agent 搭建平台，国内 \u002F 国际双版本",[606,2493,2494,2495,2496,2497],"low-code","workflow","bot-marketplace","bytedance","no-code","2026-06-24","想最快做出一个能用的 Bot，从 Coze 起步。要私有部署或开源协作，去 Dify \u002F FastGPT。","necnnd5prSTfssbiPIQOZWkZ3WDKHK3nbAKz7ZhN7uc",{"id":9,"title":10,"alternatives":2502,"api_compatible":15,"body":2503,"category":577,"chinese_friendly":566,"cover":578,"description":579,"domestic":580,"extension":581,"faq":15,"free":580,"github":558,"languages":2888,"lastVerified":584,"meta":2889,"models":15,"navigation":586,"notSuitable":15,"opensource":586,"path":587,"pillar":588,"platforms":2890,"priceTable":15,"pricing":592,"published":593,"relatedPlaybooks":15,"relatedReviews":15,"score":2891,"self_host":580,"seo":2892,"seoTitle":598,"slug":599,"sources":2893,"stem":603,"suitable":15,"tagline":604,"tags":2896,"updated":584,"verdict":611,"website":550,"__hash__":612},[12,13,14],{"type":17,"value":2504,"toc":2875},[2505,2507,2509,2511,2513,2551,2553,2591,2595,2597,2603,2619,2623,2639,2641,2667,2669,2771,2773,2805,2807,2829,2831,2835,2839,2843,2847,2849,2857,2859,2863],[20,2506,23],{"id":22},[25,2508,27],{},[25,2510,30],{},[20,2512,33],{"id":33},[35,2514,2515,2519,2523,2527,2531,2535,2539,2543,2547],{},[38,2516,2517,44],{},[41,2518,43],{},[38,2520,2521,50],{},[41,2522,49],{},[38,2524,2525,56],{},[41,2526,55],{},[38,2528,2529,62],{},[41,2530,61],{},[38,2532,2533,68],{},[41,2534,67],{},[38,2536,2537,74],{},[41,2538,73],{},[38,2540,2541,80],{},[41,2542,79],{},[38,2544,2545,86],{},[41,2546,85],{},[38,2548,2549,92],{},[41,2550,91],{},[20,2552,95],{"id":95},[97,2554,2555,2565],{},[100,2556,2557],{},[103,2558,2559,2561,2563],{},[106,2560,108],{},[106,2562,95],{},[106,2564,113],{},[115,2566,2567,2575,2583],{},[103,2568,2569,2571,2573],{},[120,2570,122],{},[120,2572,125],{},[120,2574,128],{},[103,2576,2577,2579,2581],{},[120,2578,85],{},[120,2580,135],{},[120,2582,138],{},[103,2584,2585,2587,2589],{},[120,2586,143],{},[120,2588,146],{},[120,2590,149],{},[151,2592,2593],{},[25,2594,155],{},[20,2596,159],{"id":158},[151,2598,2599],{},[25,2600,164,2601],{},[41,2602,167],{},[35,2604,2605,2607,2609,2611,2613,2615,2617],{},[38,2606,172],{},[38,2608,175],{},[38,2610,178],{},[38,2612,181],{},[38,2614,184],{},[38,2616,187],{},[38,2618,190],{},[25,2620,2621],{},[41,2622,195],{},[35,2624,2625,2627,2629,2631,2633,2635,2637],{},[38,2626,200],{},[38,2628,203],{},[38,2630,206],{},[38,2632,209],{},[38,2634,212],{},[38,2636,215],{},[38,2638,218],{},[20,2640,221],{"id":221},[223,2642,2643,2647,2651,2655,2659,2663],{},[38,2644,2645,231],{},[228,2646,230],{},[38,2648,234,2649,238],{},[228,2650,237],{},[38,2652,241,2653],{},[228,2654,244],{},[38,2656,247,2657],{},[228,2658,250],{},[38,2660,253,2661],{},[228,2662,256],{},[38,2664,259,2665,263],{},[228,2666,262],{},[20,2668,266],{"id":266},[97,2670,2671,2685],{},[100,2672,2673],{},[103,2674,2675,2677,2679,2681,2683],{},[106,2676,275],{},[106,2678,10],{},[106,2680,280],{},[106,2682,283],{},[106,2684,286],{},[115,2686,2687,2699,2711,2723,2735,2747,2759],{},[103,2688,2689,2691,2693,2695,2697],{},[120,2690,293],{},[120,2692,296],{},[120,2694,299],{},[120,2696,299],{},[120,2698,304],{},[103,2700,2701,2703,2705,2707,2709],{},[120,2702,309],{},[120,2704,312],{},[120,2706,315],{},[120,2708,318],{},[120,2710,321],{},[103,2712,2713,2715,2717,2719,2721],{},[120,2714,326],{},[120,2716,329],{},[120,2718,332],{},[120,2720,335],{},[120,2722,338],{},[103,2724,2725,2727,2729,2731,2733],{},[120,2726,343],{},[120,2728,346],{},[120,2730,349],{},[120,2732,346],{},[120,2734,354],{},[103,2736,2737,2739,2741,2743,2745],{},[120,2738,73],{},[120,2740,361],{},[120,2742,364],{},[120,2744,367],{},[120,2746,370],{},[103,2748,2749,2751,2753,2755,2757],{},[120,2750,375],{},[120,2752,378],{},[120,2754,381],{},[120,2756,381],{},[120,2758,370],{},[103,2760,2761,2763,2765,2767,2769],{},[120,2762,390],{},[120,2764,393],{},[120,2766,396],{},[120,2768,399],{},[120,2770,402],{},[20,2772,405],{"id":405},[35,2774,2775,2779,2783,2787,2791,2795,2801],{},[38,2776,2777,413],{},[41,2778,412],{},[38,2780,2781,419],{},[41,2782,418],{},[38,2784,2785,425],{},[41,2786,424],{},[38,2788,2789,431],{},[41,2790,430],{},[38,2792,2793,437],{},[41,2794,436],{},[38,2796,2797,443,2799,447],{},[41,2798,442],{},[228,2800,446],{},[38,2802,2803,453],{},[41,2804,452],{},[20,2806,457],{"id":456},[35,2808,2809,2811,2813,2815,2817,2819,2821,2823,2825,2827],{},[38,2810,462],{},[38,2812,465],{},[38,2814,468],{},[38,2816,471],{},[38,2818,474],{},[38,2820,477],{},[38,2822,480],{},[38,2824,483],{},[38,2826,486],{},[38,2828,489],{},[20,2830,493],{"id":492},[25,2832,2833,499],{},[41,2834,498],{},[25,2836,2837,505],{},[41,2838,504],{},[25,2840,2841,511],{},[41,2842,510],{},[25,2844,2845,517],{},[41,2846,516],{},[20,2848,520],{"id":520},[25,2850,2851,527,2853,527,2855],{},[524,2852,280],{"href":526},[524,2854,531],{"href":530},[524,2856,535],{"href":534},[20,2858,538],{"id":538},[151,2860,2861],{},[25,2862,543],{},[35,2864,2865,2870],{},[38,2866,2867],{},[524,2868,553],{"href":550,"rel":2869},[552],[38,2871,2872],{},[524,2873,560],{"href":558,"rel":2874},[552],{"title":562,"searchDepth":563,"depth":563,"links":2876},[2877,2878,2879,2880,2881,2882,2883,2884,2885,2886,2887],{"id":22,"depth":566,"text":23},{"id":33,"depth":566,"text":33},{"id":95,"depth":566,"text":95},{"id":158,"depth":566,"text":159},{"id":221,"depth":566,"text":221},{"id":266,"depth":566,"text":266},{"id":405,"depth":566,"text":405},{"id":456,"depth":566,"text":457},{"id":492,"depth":566,"text":493},{"id":520,"depth":566,"text":520},{"id":538,"depth":566,"text":538},[583],{},[590,591],{"power":595,"ux":563,"price":596,"cn_support":566,"stability":563},{"title":10,"description":579},[2894,2895],{"title":553,"url":550},{"title":560,"url":558},[606,607,608,609,610],{"id":2898,"title":286,"alternatives":2899,"api_compatible":15,"body":2900,"category":577,"chinese_friendly":595,"cover":3889,"description":3890,"domestic":580,"extension":581,"faq":15,"free":580,"github":3180,"languages":3891,"lastVerified":15,"meta":3893,"models":15,"navigation":586,"notSuitable":15,"opensource":586,"path":3894,"pillar":588,"platforms":3895,"priceTable":3896,"pricing":3913,"published":2463,"relatedPlaybooks":15,"relatedReviews":3914,"score":3915,"self_host":586,"seo":3916,"seoTitle":3917,"slug":618,"sources":3918,"stem":3930,"suitable":15,"tagline":3931,"tags":3932,"updated":2498,"verdict":3935,"website":3834,"__hash__":3936},"tools\u002Ftools\u002Fagent\u002Fplatform\u002Fdify.md",[2472,619,14,12],{"type":17,"value":2901,"toc":3867},[2902,2904,2922,2927,2929,2933,2936,3004,3007,3011,3014,3028,3044,3048,3051,3062,3066,3074,3085,3089,3092,3094,3098,3107,3165,3172,3176,3184,3247,3250,3270,3274,3277,3336,3342,3344,3438,3441,3470,3472,3600,3612,3647,3649,3721,3723,3725,3745,3747,3776,3778,3825,3827,3858,3863],[20,2903,23],{"id":22},[1600,2905,2907,2912],{"className":2906},[1603,1604,1605],[25,2908,2909,2911],{},[41,2910,1610],{}," Dify 是开源 LLMOps 平台的事实标准。GitHub 13 万 star、累计 100 万+ 生产 app（据 chatforest.com 2026 评测引用 Dify 官方数据），把\"可视化工作流编排 + RAG 知识库 + Agent + MCP 协议\"打包成一个 Docker Compose 能跑起来的东西。",[25,2913,1629,2914,2917,2918,2921],{},[41,2915,2916],{},"完全开源 + 模型不挑食","——同一个工作流里同时调 OpenAI、Anthropic、Ollama 本地、DeepSeek、Qwen 都行。代价是部署比 ",[524,2919,1591],{"href":2920},"\u002Fagent\u002Fplatform\u002Fcoze.html"," 折腾，新手得读 1-2 小时文档。",[151,2923,2924],{},[25,2925,2926],{},"来源说明：本文基于 docs.dify.ai 官方文档、langgenius\u002Fdify GitHub 仓库、第三方评测（besthub.dev \u002F chatforest.com \u002F joshuaopolko.com \u002F zhihu 知名专栏）综合归纳。版本号会变，部署要求请以官方最新文档为准。",[20,2928,1647],{"id":1647},[1649,2930,2932],{"id":2931},"可视化工作流chatflow-workflow","可视化工作流（Chatflow + Workflow）",[25,2934,2935],{},"Dify 把 LLM 应用拆成两种\"应用类型\"：",[97,2937,2938,2950],{},[100,2939,2940],{},[103,2941,2942,2945,2947],{},[106,2943,2944],{},"类型",[106,2946,390],{},[106,2948,2949],{},"编排范式",[115,2951,2952,2965,2978,2991],{},[103,2953,2954,2959,2962],{},[120,2955,2956],{},[41,2957,2958],{},"Chatbot",[120,2960,2961],{},"简单对话机器人",[120,2963,2964],{},"prompt + tools",[103,2966,2967,2972,2975],{},[120,2968,2969],{},[41,2970,2971],{},"Agent",[120,2973,2974],{},"自主多步任务",[120,2976,2977],{},"ReAct \u002F Function Calling",[103,2979,2980,2985,2988],{},[120,2981,2982],{},[41,2983,2984],{},"Chatflow",[120,2986,2987],{},"对话型工作流（多轮 + 分支）",[120,2989,2990],{},"节点 DAG，带聊天上下文",[103,2992,2993,2998,3001],{},[120,2994,2995],{},[41,2996,2997],{},"Workflow",[120,2999,3000],{},"单次输入→输出（API 模式）",[120,3002,3003],{},"节点 DAG，无对话状态",[25,3005,3006],{},"节点类型覆盖：LLM、知识检索、HTTP 请求、代码执行（Python \u002F JS）、条件分支、迭代、变量聚合、参数提取、问题分类——满足\"用拖拽实现可观测的 LLM pipeline\"。",[1649,3008,3010],{"id":3009},"rag-知识库","RAG 知识库",[25,3012,3013],{},"内置完整 RAG 链路：",[223,3015,3016,3019,3022,3025],{},[38,3017,3018],{},"上传文档（PDF \u002F Word \u002F Markdown \u002F 网页）",[38,3020,3021],{},"自动分块 + embedding（可配置分段策略和 embedding 模型）",[38,3023,3024],{},"混合检索（向量 + 全文 + 重排）",[38,3026,3027],{},"引用溯源（回答末尾自动附原文片段）",[25,3029,3030,3031,3036,3037,3040,3041,3043],{},"注意：根据 ",[524,3032,3035],{"href":3033,"rel":3034},"https:\u002F\u002Fzhuanlan.zhihu.com\u002Fp\u002F1887141987838309480",[552],"知乎 LLM 实战笔记 2025-03 对比"," 的实测，Dify ",[41,3038,3039],{},"社区版默认是基础语义检索","，企业版才解锁多路召回 + 重排。RAG 极致精度场景仍推荐 ",[524,3042,840],{"href":2005},"（实测准确率高 10+ 个百分点），Dify 胜在工作流而非纯 RAG。",[1649,3045,3047],{"id":3046},"模型生态40-提供商","模型生态：40+ 提供商",[25,3049,3050],{},"Dify 通过插件市场接入主流模型——OpenAI、Anthropic、Google Gemini、Azure、AWS Bedrock、Cohere、xAI、DeepSeek、Qwen、智谱、文心、豆包、月之暗面、Ollama、LM Studio、Replicate、Together AI、OpenRouter……几乎你能数出来的 LLM 提供商都在。",[25,3052,3053,3054,3056,3057,3061],{},"国产模型原生支持（不像 ",[524,3055,840],{"href":2005}," 需要 ",[524,3058,3060],{"href":3059},"\u002Fcoding\u002Fapi\u002Fone-api.html","OneAPI"," 中转），是 Dify 在国内 toB 场景流行的关键。",[1649,3063,3065],{"id":3064},"mcp-协议支持","MCP 协议支持",[25,3067,3068,3069,3073],{},"Dify 较早接入了 ",[524,3070,3072],{"href":3071},"\u002Fwiki\u002Fmcp.html","MCP（Model Context Protocol）","，工作流可以直接调 MCP Server 暴露的 tools。意味着你可以让 Dify 工作流：",[35,3075,3076,3079,3082],{},[38,3077,3078],{},"通过 MCP 调本地 PostgreSQL \u002F SQLite",[38,3080,3081],{},"通过 MCP 调 GitHub \u002F Slack \u002F Linear",[38,3083,3084],{},"通过 MCP 调自家内部系统（写一个 MCP Server 即可）",[1649,3086,3088],{"id":3087},"api-first","API-first",[25,3090,3091],{},"每个 app 自动暴露 REST API，参数和返回结构自动生成 OpenAPI Schema。集成到自家产品里不需要写包装代码，给前端 \u002F 微信小程序 \u002F 飞书机器人调用都方便。",[20,3093,1907],{"id":1907},[1649,3095,3097],{"id":3096},"云版difyai","云版（dify.ai）",[25,3099,3100,3101,3106],{},"根据 ",[524,3102,3105],{"href":3103,"rel":3104},"https:\u002F\u002Fwww.tooljunction.io\u002Fai-tools\u002Fdify-ai",[552],"tooljunction.io 2026 评测"," 引用的官方定价：",[97,3108,3109,3121],{},[100,3110,3111],{},[103,3112,3113,3116,3118],{},[106,3114,3115],{},"套餐",[106,3117,95],{},[106,3119,3120],{},"主要限制",[115,3122,3123,3134,3145,3156],{},[103,3124,3125,3128,3131],{},[120,3126,3127],{},"Sandbox",[120,3129,3130],{},"免费",[120,3132,3133],{},"200 次模型调用，1 app，5MB 知识库",[103,3135,3136,3139,3142],{},[120,3137,3138],{},"Professional",[120,3140,3141],{},"$59\u002F月起",[120,3143,3144],{},"5000 调用\u002F月，多 app，50MB 知识库",[103,3146,3147,3150,3153],{},[120,3148,3149],{},"Team",[120,3151,3152],{},"$159\u002F月起",[120,3154,3155],{},"团队协作、SSO",[103,3157,3158,3160,3162],{},[120,3159,728],{},[120,3161,146],{},[120,3163,3164],{},"定制 SLA、私有云",[25,3166,3167,3168,3171],{},"注意：云版价格只是 Dify 平台费，",[41,3169,3170],{},"模型 API 费用另算","（自带 OpenAI \u002F Anthropic key）。",[1649,3173,3175],{"id":3174},"自托管推荐","自托管（推荐）",[25,3177,3178,3183],{},[524,3179,3182],{"href":3180,"rel":3181},"https:\u002F\u002Fgithub.com\u002Flanggenius\u002Fdify",[552],"官方 GitHub 仓库"," 提供 Docker Compose 部署，社区版完全免费可商用：",[3185,3186,3190],"pre",{"className":3187,"code":3188,"language":3189,"meta":562,"style":562},"language-bash shiki shiki-themes github-light github-dark","git clone https:\u002F\u002Fgithub.com\u002Flanggenius\u002Fdify.git\ncd dify\u002Fdocker\ncp .env.example .env\ndocker compose up -d\n# 默认 http:\u002F\u002Flocalhost \u002F 端口可在 .env 调整\n","bash",[228,3191,3192,3208,3217,3228,3241],{"__ignoreMap":562},[3193,3194,3197,3201,3205],"span",{"class":3195,"line":3196},"line",1,[3193,3198,3200],{"class":3199},"sScJk","git",[3193,3202,3204],{"class":3203},"sZZnC"," clone",[3193,3206,3207],{"class":3203}," https:\u002F\u002Fgithub.com\u002Flanggenius\u002Fdify.git\n",[3193,3209,3210,3214],{"class":3195,"line":566},[3193,3211,3213],{"class":3212},"sj4cs","cd",[3193,3215,3216],{"class":3203}," dify\u002Fdocker\n",[3193,3218,3219,3222,3225],{"class":3195,"line":563},[3193,3220,3221],{"class":3199},"cp",[3193,3223,3224],{"class":3203}," .env.example",[3193,3226,3227],{"class":3203}," .env\n",[3193,3229,3230,3232,3235,3238],{"class":3195,"line":595},[3193,3231,591],{"class":3199},[3193,3233,3234],{"class":3203}," compose",[3193,3236,3237],{"class":3203}," up",[3193,3239,3240],{"class":3212}," -d\n",[3193,3242,3243],{"class":3195,"line":596},[3193,3244,3246],{"class":3245},"sJ8bj","# 默认 http:\u002F\u002Flocalhost \u002F 端口可在 .env 调整\n",[25,3248,3249],{},"硬件门槛（社区共识，非官方硬性要求）：",[35,3251,3252,3258,3264],{},[38,3253,3254,3257],{},[41,3255,3256],{},"最低","：2 核 4G，纯外接 API 模式",[38,3259,3260,3263],{},[41,3261,3262],{},"推荐","：4 核 8G + 至少 30GB 磁盘（向量数据 + 文件存储）",[38,3265,3266,3269],{},[41,3267,3268],{},"企业","：8 核 16G+，单机日活上千",[1649,3271,3273],{"id":3272},"真实-tco","真实 TCO",[25,3275,3276],{},"按一家中小团队 3 年场景估算（基于上面引用的多份评测交叉对比）：",[97,3278,3279,3292],{},[100,3280,3281],{},[103,3282,3283,3286,3289],{},[106,3284,3285],{},"成本项",[106,3287,3288],{},"云版 Professional",[106,3290,3291],{},"自托管",[115,3293,3294,3304,3314,3325],{},[103,3295,3296,3299,3302],{},[120,3297,3298],{},"平台费",[120,3300,3301],{},"~$2,100（3 年）",[120,3303,125],{},[103,3305,3306,3309,3311],{},[120,3307,3308],{},"服务器",[120,3310,125],{},[120,3312,3313],{},"~$50\u002F月 × 36 = $1,800",[103,3315,3316,3319,3322],{},[120,3317,3318],{},"模型 API",[120,3320,3321],{},"与下同",[120,3323,3324],{},"与上同",[103,3326,3327,3330,3333],{},[120,3328,3329],{},"运维人力",[120,3331,3332],{},"0",[120,3334,3335],{},"约 0.2 人月",[25,3337,3338,3341],{},[41,3339,3340],{},"结论","：日活 \u003C 100 用云版省心；> 500 或数据敏感场景自托管 ROI 更好。",[20,3343,1943],{"id":1942},[3185,3345,3347],{"className":3187,"code":3346,"language":3189,"meta":562,"style":562},"# 1. 自托管（社区版）\ngit clone https:\u002F\u002Fgithub.com\u002Flanggenius\u002Fdify.git\ncd dify\u002Fdocker\ncp .env.example .env\ndocker compose up -d\n\n# 2. 浏览器打开 http:\u002F\u002Flocalhost\n#    首次会让你创建 admin 账号\n\n# 3. 进入\"设置 → 模型供应商\"，配置 OpenAI \u002F 国产模型 API key\n\n# 4. 在主界面\"创建空白应用\"，选 Chatflow 或 Workflow\n# 5. 拖入\"开始 → LLM → 结束\"节点试一下基础 prompt\n# 6. 满意了点右上\"发布\"，自动生成 API endpoint\n",[228,3348,3349,3354,3362,3368,3376,3386,3392,3398,3404,3409,3415,3420,3426,3432],{"__ignoreMap":562},[3193,3350,3351],{"class":3195,"line":3196},[3193,3352,3353],{"class":3245},"# 1. 自托管（社区版）\n",[3193,3355,3356,3358,3360],{"class":3195,"line":566},[3193,3357,3200],{"class":3199},[3193,3359,3204],{"class":3203},[3193,3361,3207],{"class":3203},[3193,3363,3364,3366],{"class":3195,"line":563},[3193,3365,3213],{"class":3212},[3193,3367,3216],{"class":3203},[3193,3369,3370,3372,3374],{"class":3195,"line":595},[3193,3371,3221],{"class":3199},[3193,3373,3224],{"class":3203},[3193,3375,3227],{"class":3203},[3193,3377,3378,3380,3382,3384],{"class":3195,"line":596},[3193,3379,591],{"class":3199},[3193,3381,3234],{"class":3203},[3193,3383,3237],{"class":3203},[3193,3385,3240],{"class":3212},[3193,3387,3389],{"class":3195,"line":3388},6,[3193,3390,3391],{"emptyLinePlaceholder":586},"\n",[3193,3393,3395],{"class":3195,"line":3394},7,[3193,3396,3397],{"class":3245},"# 2. 浏览器打开 http:\u002F\u002Flocalhost\n",[3193,3399,3401],{"class":3195,"line":3400},8,[3193,3402,3403],{"class":3245},"#    首次会让你创建 admin 账号\n",[3193,3405,3407],{"class":3195,"line":3406},9,[3193,3408,3391],{"emptyLinePlaceholder":586},[3193,3410,3412],{"class":3195,"line":3411},10,[3193,3413,3414],{"class":3245},"# 3. 进入\"设置 → 模型供应商\"，配置 OpenAI \u002F 国产模型 API key\n",[3193,3416,3418],{"class":3195,"line":3417},11,[3193,3419,3391],{"emptyLinePlaceholder":586},[3193,3421,3423],{"class":3195,"line":3422},12,[3193,3424,3425],{"class":3245},"# 4. 在主界面\"创建空白应用\"，选 Chatflow 或 Workflow\n",[3193,3427,3429],{"class":3195,"line":3428},13,[3193,3430,3431],{"class":3245},"# 5. 拖入\"开始 → LLM → 结束\"节点试一下基础 prompt\n",[3193,3433,3435],{"class":3195,"line":3434},14,[3193,3436,3437],{"class":3245},"# 6. 满意了点右上\"发布\"，自动生成 API endpoint\n",[20,3439,3440],{"id":3440},"国内使用注意事项",[223,3442,3443,3449,3455,3461],{},[38,3444,3445,3448],{},[41,3446,3447],{},"云版 dify.ai 直连国内访问稳定但需要付款","——支持国际信用卡 \u002F Stripe",[38,3450,3451,3454],{},[41,3452,3453],{},"自托管 + 国产模型"," = 完全国内闭环，是 Dify 在国内最大优势",[38,3456,3457,3460],{},[41,3458,3459],{},"Docker 镜像拉取","：国内可能慢，建议配 Docker registry 镜像（阿里云 \u002F 网易）",[38,3462,3463,3466,3467,3469],{},[41,3464,3465],{},"数据合规","：完全自托管时，数据零外泄；某些金融 \u002F 政府客户因此从 ",[524,3468,1591],{"href":2920}," 迁到 Dify",[20,3471,1985],{"id":1985},[97,3473,3474,3494],{},[100,3475,3476],{},[103,3477,3478,3480,3482,3486,3490],{},[106,3479,275],{},[106,3481,286],{},[106,3483,3484],{},[524,3485,1591],{"href":2920},[106,3487,3488],{},[524,3489,840],{"href":2005},[106,3491,3492],{},[524,3493,2321],{"href":2320},[115,3495,3496,3509,3521,3534,3546,3559,3573,3586],{},[103,3497,3498,3500,3502,3504,3506],{},[120,3499,2018],{},[120,3501,370],{},[120,3503,381],{},[120,3505,370],{},[120,3507,3508],{},"✅（fair-code）",[103,3510,3511,3513,3515,3517,3519],{},[120,3512,2031],{},[120,3514,370],{},[120,3516,381],{},[120,3518,370],{},[120,3520,370],{},[103,3522,3523,3525,3527,3530,3532],{},[120,3524,293],{},[120,3526,2070],{},[120,3528,3529],{},"★★☆☆☆ 最简单",[120,3531,2070],{},[120,3533,2064],{},[103,3535,3536,3538,3540,3542,3544],{},[120,3537,2061],{},[120,3539,2067],{},[120,3541,2064],{},[120,3543,2070],{},[120,3545,2067],{},[103,3547,3548,3550,3552,3554,3556],{},[120,3549,878],{},[120,3551,2064],{},[120,3553,2070],{},[120,3555,2067],{},[120,3557,3558],{},"★★☆☆☆",[103,3560,3561,3564,3566,3568,3571],{},[120,3562,3563],{},"模型生态",[120,3565,2067],{},[120,3567,2064],{},[120,3569,3570],{},"★★★☆☆（OneAPI 中转）",[120,3572,2064],{},[103,3574,3575,3578,3580,3582,3584],{},[120,3576,3577],{},"中文场景",[120,3579,2064],{},[120,3581,2067],{},[120,3583,2064],{},[120,3585,2070],{},[103,3587,3588,3591,3593,3596,3598],{},[120,3589,3590],{},"字节生态绑定",[120,3592,381],{},[120,3594,3595],{},"✅（飞书\u002F抖音深度集成）",[120,3597,381],{},[120,3599,381],{},[25,3601,3602,3604,3605,2137,3608,3611],{},[41,3603,2130],{},"（基于 ",[524,3606,2136],{"href":2134,"rel":3607},[552],[524,3609,2141],{"href":1636,"rel":3610},[552]," 两份选型指南综合）：",[35,3613,3614,3620,3627,3633,3640],{},[38,3615,3616,3619],{},[41,3617,3618],{},"数据必须不出内网 + 工作流复杂"," → Dify",[38,3621,3622,2155,3625],{},[41,3623,3624],{},"个人 \u002F 小团队 \u002F 快速原型 + 字节生态",[524,3626,1591],{"href":2920},[38,3628,3629,2155,3631],{},[41,3630,2162],{},[524,3632,840],{"href":2005},[38,3634,3635,2155,3638],{},[41,3636,3637],{},"重点是连接外部 SaaS（Slack \u002F Notion \u002F 数据库）",[524,3639,2321],{"href":2320},[38,3641,3642,2155,3645],{},[41,3643,3644],{},"要画图式表达 LangChain pipeline",[524,3646,843],{"href":2301},[20,3648,2175],{"id":2175},[35,3650,3651,3657,3673,3684,3697,3703,3709,3715],{},[38,3652,3653,3656],{},[41,3654,3655],{},"社区版与企业版差距比想象大","：多路召回 \u002F 重排序 \u002F 单点登录 \u002F 审计日志都在企业版。社区版做生产前心里要有数。",[38,3658,3659,443,3665,3668,3669,3672],{},[41,3660,3661,3664],{},[228,3662,3663],{},".env"," 文件改完忘 restart",[228,3666,3667],{},"docker compose down && up -d","，不是 ",[228,3670,3671],{},"restart","——后者不重新加载 env。",[38,3674,3675,443,3678,3683],{},[41,3676,3677],{},"大版本升级会破坏数据库 schema",[524,3679,3682],{"href":3680,"rel":3681},"https:\u002F\u002Fdocs.dify.ai\u002Fzh-hans",[552],"官方升级文档"," 有详细 migration 步骤，跨大版本（如 0.x → 1.x）务必先备份 PostgreSQL 卷。生产环境强烈建议跑 staging 完整验证后再升。",[38,3685,3686,3689,3690,3692,3693,3696],{},[41,3687,3688],{},"RAG 文件大小社区版默认 15MB","：根据上述知乎实测，超过会失败。改 ",[228,3691,3663],{}," 的 ",[228,3694,3695],{},"UPLOAD_FILE_SIZE_LIMIT"," 并重启容器。",[38,3698,3699,3702],{},[41,3700,3701],{},"代码节点的 Sandbox 性能差","：内置代码执行节点跑在隔离容器里启动慢、内存小。生产高频用建议改成 HTTP 节点调外部服务。",[38,3704,3705,3708],{},[41,3706,3707],{},"工作流\"迭代节点\"循环上限","：默认 10 次，复杂 ReAct agent 容易撞天花板，需要在节点设置里调高。",[38,3710,3711,3714],{},[41,3712,3713],{},"Dify Plugin 系统是新东西","：1.0 后引入的 Plugin 体系替代了原来的 Tools\u002FModels 配置方式，老教程可能已过时——以最新官方文档为准。",[38,3716,3717,3720],{},[41,3718,3719],{},"国内 Docker 拉取镜像慢","：先配国内 registry，否则首次 pull 可能要 30+ 分钟。",[20,3722,457],{"id":456},[25,3724,2253],{},[35,3726,3727,3730,3733,3736,3739,3742],{},[38,3728,3729],{},"中大型企业 LLM 中台建设",[38,3731,3732],{},"需要私有化部署（金融 \u002F 医疗 \u002F 政府）",[38,3734,3735],{},"想做\"AI 工作流即产品\"的开发团队",[38,3737,3738],{},"同时需要 RAG + Agent + Workflow 三件套",[38,3740,3741],{},"想用国产模型 + 国际模型混合编排",[38,3743,3744],{},"已经接受 Docker + 一定运维投入",[25,3746,2273],{},[35,3748,3749,3755,3761,3764,3770],{},[38,3750,3751,3752,3754],{},"纯个人玩家做对话机器人（",[524,3753,1591],{"href":2920}," 更快）",[38,3756,3757,3758,3760],{},"只想做企业知识库 QA（",[524,3759,840],{"href":2005}," RAG 更专）",[38,3762,3763],{},"团队完全没运维能力（云版还行，自托管会踩坑）",[38,3765,3766,3767,3769],{},"需要深度对接字节飞书 \u002F 抖音（",[524,3768,1591],{"href":2920}," 原生）",[38,3771,3772,3773,3775],{},"工作流核心是连接 100+ SaaS（",[524,3774,2321],{"href":2320}," 节点更全）",[20,3777,520],{"id":520},[35,3779,3780,3790,3802,3817],{},[38,3781,2309,3782,2233,3784,2233,3786,2233,3788],{},[524,3783,1591],{"href":2920},[524,3785,840],{"href":2005},[524,3787,2321],{"href":2320},[524,3789,843],{"href":2301},[38,3791,3792,3793,2233,3795,2233,3797,2233,3800],{},"概念基础：",[524,3794,2328],{"href":2327},[524,3796,2332],{"href":2331},[524,3798,3799],{"href":3071},"MCP",[524,3801,2336],{"href":2335},[38,3803,3804,3805,2233,3809,2233,3813,2233,3815],{},"模型选型：",[524,3806,3808],{"href":3807},"\u002Fmodels\u002Fgpt-5.html","GPT-5",[524,3810,3812],{"href":3811},"\u002Fmodels\u002Fclaude-sonnet-4.html","Claude Sonnet 4",[524,3814,2350],{"href":2349},[524,3816,2358],{"href":2357},[38,3818,2361,3819,2233,3823],{},[524,3820,3822],{"href":3821},"\u002Fwiki\u002Ffine-tuning-vs-rag.html","Fine-tuning vs RAG",[524,3824,2369],{"href":2368},[20,3826,538],{"id":538},[35,3828,3829,3836,3842,3848,3855],{},[38,3830,3831,3832],{},"官网：",[524,3833,3834],{"href":3834,"rel":3835},"https:\u002F\u002Fdify.ai",[552],[38,3837,3838,3839],{},"中文文档：",[524,3840,3680],{"href":3680,"rel":3841},[552],[38,3843,3844,3845],{},"GitHub：",[524,3846,3180],{"href":3180,"rel":3847},[552],[38,3849,3850,3851],{},"官方定价：",[524,3852,3853],{"href":3853,"rel":3854},"https:\u002F\u002Fdify.ai\u002Fpricing",[552],[38,3856,3857],{},"第三方评测：tooljunction.io \u002F chatforest.com \u002F besthub.dev \u002F joshuaopolko.com \u002F 知乎 LLM 实战笔记",[25,3859,3860,3861,2406],{},"本卡片由 AIHO 编辑部根据官方公开资料与第三方评测整理。所有事实点均标注来源；如发现版本号 \u002F 价格 \u002F 功能与最新官方信息不一致，请通过 ",[524,3862,2405],{"href":2404},[3864,3865,3866],"style",{},"html pre.shiki code .sScJk, html code.shiki .sScJk{--shiki-default:#6F42C1;--shiki-dark:#B392F0}html pre.shiki code .sZZnC, html code.shiki .sZZnC{--shiki-default:#032F62;--shiki-dark:#9ECBFF}html pre.shiki code .sj4cs, html code.shiki .sj4cs{--shiki-default:#005CC5;--shiki-dark:#79B8FF}html pre.shiki code .sJ8bj, html code.shiki .sJ8bj{--shiki-default:#6A737D;--shiki-dark:#6A737D}html .default .shiki span {color: var(--shiki-default);background: var(--shiki-default-bg);font-style: var(--shiki-default-font-style);font-weight: var(--shiki-default-font-weight);text-decoration: var(--shiki-default-text-decoration);}html .shiki span {color: var(--shiki-default);background: var(--shiki-default-bg);font-style: var(--shiki-default-font-style);font-weight: var(--shiki-default-font-weight);text-decoration: var(--shiki-default-text-decoration);}html .dark .shiki span {color: var(--shiki-dark);background: var(--shiki-dark-bg);font-style: var(--shiki-dark-font-style);font-weight: var(--shiki-dark-font-weight);text-decoration: var(--shiki-dark-text-decoration);}html.dark .shiki span {color: var(--shiki-dark);background: var(--shiki-dark-bg);font-style: var(--shiki-dark-font-style);font-weight: var(--shiki-dark-font-weight);text-decoration: var(--shiki-dark-text-decoration);}",{"title":562,"searchDepth":563,"depth":563,"links":3868},[3869,3870,3877,3882,3883,3884,3885,3886,3887,3888],{"id":22,"depth":566,"text":23},{"id":1647,"depth":566,"text":1647,"children":3871},[3872,3873,3874,3875,3876],{"id":2931,"depth":563,"text":2932},{"id":3009,"depth":563,"text":3010},{"id":3046,"depth":563,"text":3047},{"id":3064,"depth":563,"text":3065},{"id":3087,"depth":563,"text":3088},{"id":1907,"depth":566,"text":1907,"children":3878},[3879,3880,3881],{"id":3096,"depth":563,"text":3097},{"id":3174,"depth":563,"text":3175},{"id":3272,"depth":563,"text":3273},{"id":1942,"depth":566,"text":1943},{"id":3440,"depth":566,"text":3440},{"id":1985,"depth":566,"text":1985},{"id":2175,"depth":566,"text":2175},{"id":456,"depth":566,"text":457},{"id":520,"depth":566,"text":520},{"id":538,"depth":566,"text":538},"\u002Fimg\u002Ftools\u002Fdify.webp","Dify 2026 真实评测：开源 LLMOps 与 AI Agent 平台，集工作流编排、RAG 知识库、Agent、MCP 和多模型接入于一体。本文对比 Coze、FastGPT、n8n，整理自托管部署、云版价格、适合团队和避坑建议。",[2426,583,3892],"ja",{},"\u002Ftools\u002Fagent\u002Fplatform\u002Fdify",[1099,1100,590,591],[3897,3901,3905,3909],{"plan":3898,"price":125,"features":3899,"notes":3900},"Self-hosted（开源版）","Docker 一键部署 + 全部核心功能（工作流 \u002F RAG \u002F Agent \u002F MCP）+ 接任意模型 API","私有部署 \u002F 完全免费 \u002F Apache 2.0",{"plan":3902,"price":125,"features":3903,"notes":3904},"Cloud Sandbox（免费云）","官方托管试水档，含基础调用配额","免运维 \u002F 试水 POC",{"plan":3906,"price":3141,"features":3907,"notes":3908},"Cloud Professional","更高调用额度 + 团队协作 + 商用支持","商用云首选",{"plan":3910,"price":3911,"features":3912,"notes":146},"Cloud Team \u002F Enterprise","Custom","更大配额 + SLA + 私有部署支持 + 合规","云版 SaaS（免费档 \u002F Professional $59\u002F月起） + 开源自托管完全免费",[2465,2466,2467,2468],{"power":596,"ux":595,"price":596,"cn_support":595,"stability":595},{"title":286,"description":3890},"Dify 评测 2026：开源 LLMOps 与 AI Agent 平台，自托管指南",[3919,3921,3923,3925,3927],{"title":3920,"url":3680},"Dify 官方文档（中文）",{"title":3922,"url":3180},"Dify GitHub",{"title":3924,"url":3853},"Dify 官方定价",{"title":3926,"url":2134},"Coze vs Dify vs FastGPT 选型",{"title":3928,"url":3929},"Dify Self-Hosted Guide 2026","https:\u002F\u002Fjoshuaopolko.com\u002Fdify-self-hosted-guide","tools\u002Fagent\u002Fplatform\u002Fdify","开源 LLMOps 平台，私有部署 Agent 首选",[606,610,1112,1113,2494,3933,3934],"llmops","mcp","想私有部署、想接全球任意模型，Dify 是答案。比 Coze 工程化、上手陡一点；比 FastGPT 工作流强、RAG 略弱。","p5aiXfjt5rD0m3qxj903DVwZMIONVWKdagLa7niYhcE",{"id":3938,"title":840,"alternatives":3939,"api_compatible":3940,"body":3942,"category":577,"chinese_friendly":596,"cover":4943,"description":4944,"domestic":580,"extension":581,"faq":15,"free":580,"github":4906,"languages":4945,"lastVerified":15,"meta":4946,"models":4947,"navigation":586,"notSuitable":4951,"opensource":586,"path":4955,"pillar":588,"platforms":4956,"priceTable":4957,"pricing":4979,"published":2463,"relatedPlaybooks":4980,"relatedReviews":4982,"score":4983,"self_host":586,"seo":4984,"seoTitle":4985,"slug":619,"sources":4986,"stem":4997,"suitable":4998,"tagline":5004,"tags":5005,"updated":2498,"verdict":5009,"website":4900,"__hash__":5010},"tools\u002Ftools\u002Fagent\u002Fplatform\u002Ffastgpt.md",[618,2472,12,14],[3941],"openai",{"type":17,"value":3943,"toc":4925},[3944,3946,3975,3986,3988,3992,3995,4009,4012,4038,4042,4049,4081,4088,4092,4145,4152,4155,4158,4161,4168,4214,4220,4224,4255,4259,4365,4368,4371,4375,4383,4504,4519,4521,4675,4683,4717,4723,4725,4799,4801,4803,4823,4825,4843,4845,4892,4894,4917,4922],[20,3945,23],{"id":22},[1600,3947,3949,3964],{"className":3948},[1603,1604,1605],[25,3950,3951,3953,3954,3959,3960,3963],{},[41,3952,1610],{}," labring 团队开源的 LLM 知识库 RAG 平台，27k+ GitHub star（截至 2026-03 数据，",[524,3955,3958],{"href":3956,"rel":3957},"https:\u002F\u002Fcloud.tencent.com\u002Fdeveloper\u002Farticle\u002F2632669",[552],"腾讯云 2026-03 教程"," 引用），Apache 2.0 许可证可商用。",[41,3961,3962],{},"核心优势是 RAG 链路工程做得极细","——问题预处理、混合检索、重排序、上下文组装、答案生成每一步都可视化调参。",[25,3965,3966,3967,3970,3971,3974],{},"最大价值在 ",[41,3968,3969],{},"国内企业知识库 + 私有部署"," 场景。代价是 ",[41,3972,3973],{},"配置门槛","：docker 基础 + 网络知识 + 一定运维能力。",[151,3976,3977],{},[25,3978,3979,3980,3985],{},"来源说明：本文基于 fastgpt.io 官方页面、github.com\u002Flabring\u002FFastGPT 仓库、",[524,3981,3984],{"href":3982,"rel":3983},"https:\u002F\u002Fwww.nanhuantech.com\u002Fzh\u002Fai-reviews\u002Ffastgpt-2025-review",[552],"南环 AI 2026-05 评测","、腾讯云开发者社区 2026-03 部署教程综合整理。版本迭代较快，命令和价格请以最新官方文档为准。",[20,3987,1647],{"id":1647},[1649,3989,3991],{"id":3990},"知识库管理核心能力","知识库管理（核心能力）",[25,3993,3994],{},"支持文件类型：",[35,3996,3997,4000,4003,4006],{},[38,3998,3999],{},"文档：PDF \u002F Word \u002F Markdown \u002F TXT \u002F HTML",[38,4001,4002],{},"表格：Excel \u002F CSV",[38,4004,4005],{},"网页：URL 抓取 + 定时同步",[38,4007,4008],{},"API：通过接口推送内容",[25,4010,4011],{},"处理流程：上传 → 文本切分 → 向量化 → 存储 → 可用于问答。支持：",[35,4013,4014,4020,4026,4032],{},[38,4015,4016,4019],{},[41,4017,4018],{},"文件夹分组","：不同主题 \u002F 部门分类",[38,4021,4022,4025],{},[41,4023,4024],{},"多种分块策略","：默认按段落 \u002F 按 token 数 \u002F 自定义",[38,4027,4028,4031],{},[41,4029,4030],{},"批量导入","：脚本化大批量同步",[38,4033,4034,4037],{},[41,4035,4036],{},"定时同步","：网页源自动更新",[1649,4039,4041],{"id":4040},"rag-流程编排最强卖点","RAG 流程编排（最强卖点）",[25,4043,4044,4048],{},[524,4045,4047],{"href":3982,"rel":4046},[552],"南环 AI 2026 评测"," 总结的 FastGPT RAG 链路：",[223,4050,4051,4057,4063,4069,4075],{},[38,4052,4053,4056],{},[41,4054,4055],{},"问题预处理","：改写 \u002F 扩展 \u002F 错词纠正（提升召回率）",[38,4058,4059,4062],{},[41,4060,4061],{},"检索策略","：语义检索 \u002F 关键词 BM25 \u002F 混合检索，可调相似度阈值",[38,4064,4065,4068],{},[41,4066,4067],{},"重排序（Rerank）","：对初步检索结果二次排序，提升相关性",[38,4070,4071,4074],{},[41,4072,4073],{},"上下文组装","：最优 chunk + 问题 → prompt",[38,4076,4077,4080],{},[41,4078,4079],{},"答案生成","：调大模型基于检索结果回答 + 引用标注",[25,4082,4083,4084,4087],{},"每一步都可视化调参，这是 FastGPT 比 Coze \u002F Dify 在 ",[41,4085,4086],{},"纯知识库 QA 精度","上更高的原因。",[1649,4089,4091],{"id":4090},"多模型支持不绑定厂商","多模型支持（不绑定厂商）",[97,4093,4094,4104],{},[100,4095,4096],{},[103,4097,4098,4101],{},[106,4099,4100],{},"模型类别",[106,4102,4103],{},"支持",[115,4105,4106,4114,4121,4129,4137],{},[103,4107,4108,4111],{},[120,4109,4110],{},"国产闭源",[120,4112,4113],{},"豆包 \u002F 通义千问 \u002F 文心一言 \u002F 智谱 GLM \u002F Moonshot Kimi \u002F MiniMax",[103,4115,4116,4118],{},[120,4117,2018],{},[120,4119,4120],{},"LLaMA \u002F Qwen \u002F ChatGLM \u002F DeepSeek 等可自部署",[103,4122,4123,4126],{},[120,4124,4125],{},"OpenAI 系",[120,4127,4128],{},"GPT-5 \u002F GPT-5 mini \u002F o3",[103,4130,4131,4134],{},[120,4132,4133],{},"Claude 系",[120,4135,4136],{},"Sonnet 4 \u002F Opus 4 \u002F Haiku",[103,4138,4139,4142],{},[120,4140,4141],{},"嵌入 \u002F 重排",[120,4143,4144],{},"BGE \u002F m3e \u002F OpenAI text-embedding-3",[25,4146,4147,4148,4151],{},"可以在 ",[41,4149,4150],{},"应用级别","为不同知识库 \u002F 不同场景配置不同模型，做\"低成本 embedding + 高质量 LLM 生成\"组合。",[1649,4153,4154],{"id":4154},"工作流与高级编排",[25,4156,4157],{},"新版本（v4.14.x）支持类似 Dify 的工作流节点编排——条件分支、循环、HTTP 调用、代码节点。能做\"分类 → 路由到不同子知识库 → 不同模型回答\"这类复杂场景。",[1649,4159,4160],{"id":4160},"多向量库选择",[25,4162,4163,4167],{},[524,4164,4166],{"href":3956,"rel":4165},[552],"腾讯云教程"," 公开的 4 种向量后端：",[97,4169,4170,4180],{},[100,4171,4172],{},[103,4173,4174,4177],{},[106,4175,4176],{},"后端",[106,4178,4179],{},"适用",[115,4181,4182,4190,4198,4206],{},[103,4183,4184,4187],{},[120,4185,4186],{},"PgVector",[120,4188,4189],{},"5000 万索引以下，新手 \u002F 小规模",[103,4191,4192,4195],{},[120,4193,4194],{},"Milvus",[120,4196,4197],{},"亿级以上，高性能",[103,4199,4200,4203],{},[120,4201,4202],{},"Zilliz Cloud",[120,4204,4205],{},"Milvus 全托管 SaaS",[103,4207,4208,4211],{},[120,4209,4210],{},"SeekDB \u002F OceanBase",[120,4212,4213],{},"企业级国产化",[25,4215,4216,4217,1904],{},"部署时选对应 ",[228,4218,4219],{},"docker-compose.{pgvector|milvus|...}.yml",[1649,4221,4223],{"id":4222},"api-与-mcp","API 与 MCP",[35,4225,4226,4232,4238,4249],{},[38,4227,4228,4231],{},[41,4229,4230],{},"对话 API","：流式 \u002F 非流式 HTTP，OpenAI 兼容",[38,4233,4234,4237],{},[41,4235,4236],{},"知识库检索 API","：单独调检索（不走生成）做 hybrid pipeline",[38,4239,4240,4243,4244,4248],{},[41,4241,4242],{},"MCP Server","：3005 端口暴露 MCP SSE 服务，可被 ",[524,4245,4247],{"href":4246},"\u002Fcoding\u002Fcli\u002Fclaude-code.html","Claude Code"," 等客户端直接接入",[38,4250,4251,4254],{},[41,4252,4253],{},"Webhook","：回调通知",[20,4256,4258],{"id":4257},"部署-10-分钟docker","部署 10 分钟（Docker）",[3185,4260,4262],{"className":3187,"code":4261,"language":3189,"meta":562,"style":562},"# 克隆代码\ngit clone https:\u002F\u002Fgithub.com\u002Flabring\u002FFastGPT.git\ncd FastGPT\n\n# 切到最新稳定版（参考 GitHub releases）\ngit switch -c 4.14.7.2\n\n# 选向量库版本（个人 \u002F 小规模选 pg）\ncd deploy\u002Fdocker\u002Fcn\nwget https:\u002F\u002Fdoc.fastgpt.cn\u002Fdeploy\u002Fconfig\u002Fconfig.json\n\n# 启动\ndocker-compose -f docker-compose.pg.yml up -d\n\n# 访问 http:\u002F\u002F\u003Cip>:3000，默认账号 root \u002F 1234\n",[228,4263,4264,4269,4278,4285,4289,4294,4307,4311,4316,4323,4331,4335,4340,4355,4359],{"__ignoreMap":562},[3193,4265,4266],{"class":3195,"line":3196},[3193,4267,4268],{"class":3245},"# 克隆代码\n",[3193,4270,4271,4273,4275],{"class":3195,"line":566},[3193,4272,3200],{"class":3199},[3193,4274,3204],{"class":3203},[3193,4276,4277],{"class":3203}," https:\u002F\u002Fgithub.com\u002Flabring\u002FFastGPT.git\n",[3193,4279,4280,4282],{"class":3195,"line":563},[3193,4281,3213],{"class":3212},[3193,4283,4284],{"class":3203}," FastGPT\n",[3193,4286,4287],{"class":3195,"line":595},[3193,4288,3391],{"emptyLinePlaceholder":586},[3193,4290,4291],{"class":3195,"line":596},[3193,4292,4293],{"class":3245},"# 切到最新稳定版（参考 GitHub releases）\n",[3193,4295,4296,4298,4301,4304],{"class":3195,"line":3388},[3193,4297,3200],{"class":3199},[3193,4299,4300],{"class":3203}," switch",[3193,4302,4303],{"class":3212}," -c",[3193,4305,4306],{"class":3212}," 4.14.7.2\n",[3193,4308,4309],{"class":3195,"line":3394},[3193,4310,3391],{"emptyLinePlaceholder":586},[3193,4312,4313],{"class":3195,"line":3400},[3193,4314,4315],{"class":3245},"# 选向量库版本（个人 \u002F 小规模选 pg）\n",[3193,4317,4318,4320],{"class":3195,"line":3406},[3193,4319,3213],{"class":3212},[3193,4321,4322],{"class":3203}," deploy\u002Fdocker\u002Fcn\n",[3193,4324,4325,4328],{"class":3195,"line":3411},[3193,4326,4327],{"class":3199},"wget",[3193,4329,4330],{"class":3203}," https:\u002F\u002Fdoc.fastgpt.cn\u002Fdeploy\u002Fconfig\u002Fconfig.json\n",[3193,4332,4333],{"class":3195,"line":3417},[3193,4334,3391],{"emptyLinePlaceholder":586},[3193,4336,4337],{"class":3195,"line":3422},[3193,4338,4339],{"class":3245},"# 启动\n",[3193,4341,4342,4345,4348,4351,4353],{"class":3195,"line":3428},[3193,4343,4344],{"class":3199},"docker-compose",[3193,4346,4347],{"class":3212}," -f",[3193,4349,4350],{"class":3203}," docker-compose.pg.yml",[3193,4352,3237],{"class":3203},[3193,4354,3240],{"class":3212},[3193,4356,4357],{"class":3195,"line":3434},[3193,4358,3391],{"emptyLinePlaceholder":586},[3193,4360,4362],{"class":3195,"line":4361},15,[3193,4363,4364],{"class":3245},"# 访问 http:\u002F\u002F\u003Cip>:3000，默认账号 root \u002F 1234\n",[25,4366,4367],{},"最低配置：2C4G + 20GB 硬盘 + Docker 28+ + Docker Compose 2.20+。",[25,4369,4370],{},"进入后台 → 账号 → 模型提供商 → 配置至少 1 个对话模型 + 1 个嵌入模型 → 即可开始建知识库。",[20,4372,4374],{"id":4373},"云版-vs-自托管对比","云版 vs 自托管对比",[25,4376,4377,4382],{},[524,4378,4381],{"href":4379,"rel":4380},"https:\u002F\u002Ffastgpt.io\u002Fzh\u002Fprice",[552],"fastgpt.io 官方定价"," 公开数据：",[97,4384,4385,4410],{},[100,4386,4387],{},[103,4388,4389,4391,4393,4396,4399,4402,4404,4407],{},[106,4390,3115],{},[106,4392,95],{},[106,4394,4395],{},"AI 积分",[106,4397,4398],{},"知识库索引",[106,4400,4401],{},"团队",[106,4403,2971],{},[106,4405,4406],{},"知识库",[106,4408,4409],{},"QPM",[115,4411,4412,4436,4460,4484],{},[103,4413,4414,4416,4418,4421,4424,4427,4430,4433],{},[120,4415,3130],{},[120,4417,2444],{},[120,4419,4420],{},"100",[120,4422,4423],{},"600",[120,4425,4426],{},"1",[120,4428,4429],{},"10",[120,4431,4432],{},"3",[120,4434,4435],{},"30",[103,4437,4438,4440,4443,4446,4449,4452,4455,4457],{},[120,4439,894],{},[120,4441,4442],{},"¥99\u002F月",[120,4444,4445],{},"4000",[120,4447,4448],{},"6000",[120,4450,4451],{},"5",[120,4453,4454],{},"50",[120,4456,4435],{},[120,4458,4459],{},"300",[103,4461,4462,4465,4468,4471,4474,4476,4479,4481],{},[120,4463,4464],{},"高级",[120,4466,4467],{},"¥599\u002F月",[120,4469,4470],{},"25000",[120,4472,4473],{},"36000",[120,4475,4454],{},[120,4477,4478],{},"200",[120,4480,4420],{},[120,4482,4483],{},"1500",[103,4485,4486,4489,4491,4494,4496,4498,4500,4502],{},[120,4487,4488],{},"定制",[120,4490,2454],{},[120,4492,4493],{},"弹性",[120,4495,4493],{},[120,4497,4493],{},[120,4499,4493],{},[120,4501,4493],{},[120,4503,4493],{},[25,4505,4506,4509,4510,4513,4514,4518],{},[41,4507,4508],{},"云版适合","：不想运维、量小、要快速上线\n",[41,4511,4512],{},"自托管适合","：量大（10 万+ 日问答）、数据敏感、要深度定制——按 ",[524,4515,4517],{"href":3982,"rel":4516},[552],"南环评测"," 估算：\"日均 10 万次问答的企业场景，商业 SaaS 年费数十万，自建 FastGPT + 开源模型只需数万硬件投入\"",[20,4520,1985],{"id":1985},[97,4522,4523,4543],{},[100,4524,4525],{},[103,4526,4527,4529,4531,4535,4539,4541],{},[106,4528,275],{},[106,4530,840],{},[106,4532,4533],{},[524,4534,286],{"href":2000},[106,4536,4537],{},[524,4538,1591],{"href":2920},[106,4540,1053],{},[106,4542,616],{},[115,4544,4545,4565,4581,4596,4613,4628,4644,4659],{},[103,4546,4547,4550,4553,4556,4559,4562],{},[120,4548,4549],{},"核心定位",[120,4551,4552],{},"知识库 QA",[120,4554,4555],{},"综合 LLMOps",[120,4557,4558],{},"Bot + 工作流",[120,4560,4561],{},"文档解析+RAG",[120,4563,4564],{},"桌面级 KB",[103,4566,4567,4569,4572,4574,4576,4578],{},[120,4568,2018],{},[120,4570,4571],{},"✅ Apache 2.0",[120,4573,4571],{},[120,4575,381],{},[120,4577,4571],{},[120,4579,4580],{},"✅ MIT",[103,4582,4583,4585,4588,4590,4592,4594],{},[120,4584,2031],{},[120,4586,4587],{},"★★★★★ docker",[120,4589,2067],{},[120,4591,2034],{},[120,4593,2064],{},[120,4595,2067],{},[103,4597,4598,4601,4604,4606,4608,4611],{},[120,4599,4600],{},"RAG 深度",[120,4602,4603],{},"★★★★★ 最细",[120,4605,2064],{},[120,4607,2070],{},[120,4609,4610],{},"★★★★★ 文档解析最强",[120,4612,2070],{},[103,4614,4615,4618,4620,4622,4624,4626],{},[120,4616,4617],{},"工作流",[120,4619,2064],{},[120,4621,2067],{},[120,4623,2064],{},[120,4625,2070],{},[120,4627,3558],{},[103,4629,4630,4632,4635,4637,4640,4642],{},[120,4631,221],{},[120,4633,4634],{},"★★★☆☆ 需 docker",[120,4636,2064],{},[120,4638,4639],{},"★★★★★ 最简单",[120,4641,2070],{},[120,4643,2064],{},[103,4645,4646,4649,4651,4653,4655,4657],{},[120,4647,4648],{},"中文优化",[120,4650,2067],{},[120,4652,2064],{},[120,4654,2067],{},[120,4656,2064],{},[120,4658,2070],{},[103,4660,4661,4664,4667,4669,4671,4673],{},[120,4662,4663],{},"多平台发布",[120,4665,4666],{},"⚠️ API 为主",[120,4668,2064],{},[120,4670,2067],{},[120,4672,2041],{},[120,4674,2041],{},[25,4676,4677,2131,4679,1678],{},[41,4678,2130],{},[524,4680,4682],{"href":3982,"rel":4681},[552],"南环 AI 评测",[35,4684,4685,4691,4698,4705,4711],{},[38,4686,4687,4690],{},[41,4688,4689],{},"核心需求是 RAG 精度"," → FastGPT",[38,4692,4693,2155,4696],{},[41,4694,4695],{},"需要丰富插件 + 复杂工作流 + 多平台发布",[524,4697,286],{"href":2000},[38,4699,4700,2155,4703],{},[41,4701,4702],{},"零代码、快速发布到飞书 \u002F 微信",[524,4704,1591],{"href":2920},[38,4706,4707,4710],{},[41,4708,4709],{},"文档解析（含 OCR \u002F 表格 \u002F 公式）是瓶颈"," → RAGFlow",[38,4712,4713,4716],{},[41,4714,4715],{},"桌面 \u002F 单机使用"," → AnythingLLM",[25,4718,4719,4722],{},[41,4720,4721],{},"很多企业同时用","：FastGPT 做知识库底座 + Coze 做前端 Bot 发布 \u002F 工作流编排。",[20,4724,2175],{"id":2175},[35,4726,4727,4740,4749,4755,4765,4777,4783,4789],{},[38,4728,4729,443,4732,4735,4736,4739],{},[41,4730,4731],{},"docker-compose 镜像 tag 不一致",[524,4733,4166],{"href":3956,"rel":4734},[552]," 实测的坑——某些版本编排文件的 image tag 与最新 release 不一致，启动报\"镜像找不到\"，手动改 ",[228,4737,4738],{},"image:"," 行为正确版本即可",[38,4741,4742,4745,4746,4748],{},[41,4743,4744],{},"3000 端口冲突","：默认占用 3000（主服务）\u002F 9000（S3 \u002F MinIO）\u002F 3005（MCP）；改 ",[228,4747,4344],{}," 的 ports 映射端口",[38,4750,4751,4754],{},[41,4752,4753],{},"PostgreSQL pgvector 不够用就换 Milvus","：单库索引超 5000 万时 pgvector 查询性能下降，切 Milvus",[38,4756,4757,4760,4761,4764],{},[41,4758,4759],{},"向量库选错代价大","：先评估索引量再选向量后端，迁移要重新 embedding 整库，按 ",[524,4762,4517],{"href":3982,"rel":4763},[552],"：\"新手 \u002F 小规模 PgVector，中大规模 Milvus，企业 \u002F 国产 OceanBase\"",[38,4766,4767,443,4770,4773,4774],{},[41,4768,4769],{},"MinIO 默认密码",[228,4771,4772],{},"minioadmin\u002Fminioadmin","，",[41,4775,4776],{},"部署到公网前必须改",[38,4778,4779,4782],{},[41,4780,4781],{},"分段策略影响巨大","：默认分段对长法律 \u002F 医疗文档不友好，需调\"按章节\"或\"自定义\"",[38,4784,4785,4788],{},[41,4786,4787],{},"嵌入模型 ≠ 对话模型","：经常有人只配 GPT-4 没配 embedding 模型，知识库无法索引——必须同时配两类",[38,4790,4791,4794,4795,2208],{},[41,4792,4793],{},"云版 AI 积分会过期","：未用完不能跨月累积（按 ",[524,4796,4798],{"href":4379,"rel":4797},[552],"fastgpt.io 定价 FAQ",[20,4800,457],{"id":456},[25,4802,2253],{},[35,4804,4805,4808,4811,4814,4817,4820],{},[38,4806,4807],{},"企业内部知识库（员工手册 \u002F 制度 \u002F 流程）",[38,4809,4810],{},"产品 FAQ \u002F 用户手册问答",[38,4812,4813],{},"医疗 \u002F 法律 \u002F 金融垂直领域知识系统",[38,4815,4816],{},"数据严格不出网 + Apache 2.0 商用",[38,4818,4819],{},"有 docker 运维基础的技术团队",[38,4821,4822],{},"需要把 RAG 当后端服务的开发者（API 接入业务系统）",[25,4824,2273],{},[35,4826,4827,4832,4837,4840],{},[38,4828,4829,4830,2208],{},"完全非技术用户（去 ",[524,4831,1591],{"href":2920},[38,4833,4834,4835,2208],{},"主要需求是工作流 + 插件集成（去 ",[524,4836,286],{"href":2000},[38,4838,4839],{},"文档解析 \u002F OCR 是首要痛点（RAGFlow）",[38,4841,4842],{},"不想自己运维 + 量很小（FastGPT 云免费版起步即可）",[20,4844,520],{"id":520},[35,4846,4847,4856,4872,4886],{},[38,4848,2309,4849,2233,4851,4853,4854],{},[524,4850,286],{"href":2000},[524,4852,1591],{"href":2920}," \u002F RAGFlow \u002F AnythingLLM \u002F ",[524,4855,2321],{"href":2320},[38,4857,2324,4858,2233,4860,2233,4864,2233,4867,2233,4870],{},[524,4859,2332],{"href":2331},[524,4861,4863],{"href":4862},"\u002Fwiki\u002Fembedding.html","Embedding",[524,4865,4866],{"href":4862},"Vector Database",[524,4868,4869],{"href":2331},"Reranker",[524,4871,2328],{"href":2327},[38,4873,2342,4874,2233,4876,2233,4878,2233,4880,2233,4884],{},[524,4875,2350],{"href":2349},[524,4877,2354],{"href":2353},[524,4879,2358],{"href":2357},[524,4881,4883],{"href":4882},"\u002Fmodels\u002Fkimi-k2.html","Kimi K2",[524,4885,2346],{"href":2345},[38,4887,2361,4888,2233,4890],{},[524,4889,2369],{"href":2368},[524,4891,2365],{"href":2364},[20,4893,538],{"id":538},[35,4895,4896,4902,4908,4914],{},[38,4897,3831,4898],{},[524,4899,4900],{"href":4900,"rel":4901},"https:\u002F\u002Ffastgpt.io",[552],[38,4903,3844,4904],{},[524,4905,4906],{"href":4906,"rel":4907},"https:\u002F\u002Fgithub.com\u002Flabring\u002FFastGPT",[552],[38,4909,4910,4911],{},"定价：",[524,4912,4379],{"href":4379,"rel":4913},[552],[38,4915,4916],{},"第三方评测：南环 AI \u002F 腾讯云开发者社区 \u002F 飞书 AGI 掘金知识库",[25,4918,4919,4920,2406],{},"本卡片由 AIHO 编辑部根据官方公开资料与第三方评测整理。所有事实点均标注来源；如发现价格 \u002F 命令 \u002F 功能与最新官方信息不一致，请通过 ",[524,4921,2405],{"href":2404},[3864,4923,4924],{},"html pre.shiki code .sJ8bj, html code.shiki .sJ8bj{--shiki-default:#6A737D;--shiki-dark:#6A737D}html pre.shiki code .sScJk, html code.shiki .sScJk{--shiki-default:#6F42C1;--shiki-dark:#B392F0}html pre.shiki code .sZZnC, html code.shiki .sZZnC{--shiki-default:#032F62;--shiki-dark:#9ECBFF}html pre.shiki code .sj4cs, html code.shiki .sj4cs{--shiki-default:#005CC5;--shiki-dark:#79B8FF}html .default .shiki span {color: var(--shiki-default);background: var(--shiki-default-bg);font-style: var(--shiki-default-font-style);font-weight: var(--shiki-default-font-weight);text-decoration: var(--shiki-default-text-decoration);}html .shiki span {color: var(--shiki-default);background: var(--shiki-default-bg);font-style: var(--shiki-default-font-style);font-weight: var(--shiki-default-font-weight);text-decoration: var(--shiki-default-text-decoration);}html .dark .shiki span {color: var(--shiki-dark);background: var(--shiki-dark-bg);font-style: var(--shiki-dark-font-style);font-weight: var(--shiki-dark-font-weight);text-decoration: var(--shiki-dark-text-decoration);}html.dark .shiki span {color: var(--shiki-dark);background: var(--shiki-dark-bg);font-style: var(--shiki-dark-font-style);font-weight: var(--shiki-dark-font-weight);text-decoration: var(--shiki-dark-text-decoration);}",{"title":562,"searchDepth":563,"depth":563,"links":4926},[4927,4928,4936,4937,4938,4939,4940,4941,4942],{"id":22,"depth":566,"text":23},{"id":1647,"depth":566,"text":1647,"children":4929},[4930,4931,4932,4933,4934,4935],{"id":3990,"depth":563,"text":3991},{"id":4040,"depth":563,"text":4041},{"id":4090,"depth":563,"text":4091},{"id":4154,"depth":563,"text":4154},{"id":4160,"depth":563,"text":4160},{"id":4222,"depth":563,"text":4223},{"id":4257,"depth":566,"text":4258},{"id":4373,"depth":566,"text":4374},{"id":1985,"depth":566,"text":1985},{"id":2175,"depth":566,"text":2175},{"id":456,"depth":566,"text":457},{"id":520,"depth":566,"text":520},{"id":538,"depth":566,"text":538},"\u002Fimg\u002Ftools\u002Ffastgpt.webp","FastGPT 真实评测：开源 LLM 知识库 RAG 平台，labring 团队出品，27k+ GitHub star。一键 docker-compose 部署、RAG 流程编排可视化、多向量库支持。AIHO 编辑部基于官方文档与社区资料整理，含与 Dify\u002FCoze 对比、避坑指南。",[2426,583],{},[2430,2431,2429,4948,4949,4950],"gpt-4o","claude-sonnet-4","kimi",[4952,4953,4954],"完全零代码 \u002F 不懂 docker 的用户（去 Coze）","Bot 多平台一键发布场景（Coze 强项）","插件 \u002F 工作流复杂集成（去 Dify）","\u002Ftools\u002Fagent\u002Fplatform\u002Ffastgpt",[1099,1100,590],[4958,4962,4967,4971,4975],{"plan":4959,"price":3130,"limit":4960,"cn_pay":2446,"note":4961},"Self-host 开源","全功能 + 全数据本地","Apache 2.0 可商用",{"plan":4963,"price":4964,"limit":4965,"cn_pay":2446,"note":4966},"云免费版","¥0\u002F月","100 AI 积分 + 600 索引 + 3 知识库","试水",{"plan":4968,"price":4442,"limit":4969,"cn_pay":2451,"note":4970},"云基础版","4000 积分 + 6000 索引 + 50 Agent","中小团队 SaaS",{"plan":4972,"price":4467,"limit":4973,"cn_pay":370,"note":4974},"云高级版","25000 积分 + 36000 索引 + 50 成员 + 200 Agent + 1500 QPM","企业级生产",{"plan":4976,"price":2454,"limit":4977,"cn_pay":370,"note":4978},"云定制版","弹性资源 + 深度技术支持 + 专属客户经理","中大型企业","自托管开源免费 \u002F 云版 ¥0-¥599\u002F月",[4981],"onboarding\u002Ffastgpt-getting-started",[2468,2465,2466,2467],{"power":595,"ux":595,"price":596,"cn_support":596,"stability":595},{"title":840,"description":4944},"FastGPT 评测 2026：开源知识库问答平台，AI 工作流引擎，对比 Dify",[4987,4989,4991,4993,4995],{"title":4988,"url":4900},"FastGPT 官网",{"title":4990,"url":4906},"FastGPT GitHub",{"title":4992,"url":4379},"FastGPT 定价页",{"title":4994,"url":3982},"FastGPT 2025 测评（南环 AI）",{"title":4996,"url":3956},"FastGPT 部署教程（腾讯云）","tools\u002Fagent\u002Fplatform\u002Ffastgpt",[4999,5000,5001,5002,5003],"企业内部知识库（员工手册、规章、流程）","产品文档智能问答（FAQ \u002F 用户手册）","垂直领域知识库（医疗、法律、金融）","数据严格不出网的合规场景","需要精细 RAG 流程编排（重排序、混合检索、阈值调节）","开源知识库问答系统，国内私有部署友好",[606,610,1112,1113,5006,5007,5008],"china","knowledge-base","labring","国内企业知识库私有化首选。RAG 召回工程做得很细，可视化调试好用，docker-compose 一键部署。生态插件不如 Dify 丰富。","NAay3javdz1FV9ZaZVXCpoCFdZlJu5B4E0Y3oSzmYLk",{"id":5012,"title":531,"alternatives":5013,"api_compatible":15,"body":5014,"category":577,"chinese_friendly":563,"cover":5510,"description":5511,"domestic":580,"extension":581,"faq":15,"free":580,"github":5495,"languages":5512,"lastVerified":584,"meta":5513,"models":15,"navigation":586,"notSuitable":15,"opensource":586,"path":5514,"pillar":588,"platforms":5515,"priceTable":15,"pricing":5516,"published":593,"relatedPlaybooks":15,"relatedReviews":15,"score":5517,"self_host":580,"seo":5518,"seoTitle":5519,"slug":5520,"sources":5521,"stem":5524,"suitable":15,"tagline":5525,"tags":5526,"updated":584,"verdict":5529,"website":5489,"__hash__":5530},"tools\u002Ftools\u002Fagent\u002Fplatform\u002Fflowise.md",[12,618,14],{"type":17,"value":5015,"toc":5497},[5016,5018,5021,5024,5026,5081,5083,5137,5141,5143,5149,5172,5176,5199,5201,5225,5227,5358,5360,5404,5406,5438,5440,5446,5452,5461,5467,5469,5477,5479,5483],[20,5017,23],{"id":22},[25,5019,5020],{},"Flowise 是开源 LLM 流程编排平台（Apache 2.0），拖拽式可视化构建 AI 应用，底层基于 LangChain \u002F LlamaIndex 生态。节点包括 LLM \u002F Chat Model \u002F Embedding \u002F Vector Store \u002F Tool \u002F Agent \u002F Memory 等，连线编排成完整流程。Docker 自托管 + Cloud 云端双模式，导出为 API \u002F 嵌入式聊天组件 \u002F SDK。",[25,5022,5023],{},"适合：不写代码也要搭建 AI 应用的产品\u002F运营团队、快速验证 RAG \u002F Chatbot 原型、LangChain 生态用户可视化探索。不适合：复杂业务逻辑编排（用 n8n \u002F Dify）、大规模并发生产服务、深度定制需求（直接写 LangChain 代码）。",[20,5025,33],{"id":33},[35,5027,5028,5034,5040,5046,5052,5058,5063,5069,5075],{},[38,5029,5030,5033],{},[41,5031,5032],{},"可视化拖拽编排","：节点 + 连线构建 LLM 流程，实时预览",[38,5035,5036,5039],{},[41,5037,5038],{},"LangChain 生态","：直接使用 LangChain \u002F LlamaIndex 全部组件",[38,5041,5042,5045],{},[41,5043,5044],{},"丰富节点","：LLM \u002F Chat Model \u002F Embedding \u002F Vector Store \u002F Tool \u002F Agent \u002F Memory \u002F Chain",[38,5047,5048,5051],{},[41,5049,5050],{},"Agent 支持","：Conversational Agent \u002F Tool Calling Agent \u002F ReAct Agent",[38,5053,5054,5057],{},[41,5055,5056],{},"RAG 流程","：文档加载 → 切片 → 嵌入 → 向量存储 → 检索 → 生成，全可视化",[38,5059,5060,5062],{},[41,5061,662],{},"：Pinecone \u002F Qdrant \u002F Chroma \u002F Weaviate \u002F Supabase",[38,5064,5065,5068],{},[41,5066,5067],{},"多模型接入","：OpenAI \u002F Claude \u002F Gemini \u002F Azure \u002F Ollama \u002F HuggingFace",[38,5070,5071,5074],{},[41,5072,5073],{},"部署方式","：导出 REST API \u002F 嵌入式聊天组件 \u002F React SDK",[38,5076,5077,5080],{},[41,5078,5079],{},"凭据管理","：API Key 加密存储，支持环境变量",[20,5082,95],{"id":95},[97,5084,5085,5095],{},[100,5086,5087],{},[103,5088,5089,5091,5093],{},[106,5090,108],{},[106,5092,95],{},[106,5094,113],{},[115,5096,5097,5106,5117,5128],{},[103,5098,5099,5101,5103],{},[120,5100,122],{},[120,5102,125],{},[120,5104,5105],{},"完整功能，Apache 2.0，自托管",[103,5107,5108,5111,5114],{},[120,5109,5110],{},"Cloud Starter",[120,5112,5113],{},"$39\u002F月起",[120,5115,5116],{},"托管服务，1 个工作区",[103,5118,5119,5122,5125],{},[120,5120,5121],{},"Cloud Pro",[120,5123,5124],{},"$89\u002F月起",[120,5126,5127],{},"多工作区 + 团队协作 + 高并发",[103,5129,5130,5132,5134],{},[120,5131,728],{},[120,5133,146],{},[120,5135,5136],{},"私有部署 + SSO + 专属支持",[151,5138,5139],{},[25,5140,155],{},[20,5142,159],{"id":158},[151,5144,5145],{},[25,5146,164,5147],{},[41,5148,167],{},[35,5150,5151,5154,5157,5160,5166,5169],{},[38,5152,5153],{},"拖拽编排体验流畅，LangChain 组件全覆盖，不用写代码也能搭复杂流程",[38,5155,5156],{},"RAG 流程模板开箱即用，上传 PDF + 接 OpenAI 几分钟出问答机器人",[38,5158,5159],{},"嵌入式聊天组件方便，生成一段 JS 代码嵌入网页即可",[38,5161,5162,5163,5165],{},"Docker 部署简单，",[228,5164,753],{}," 一条命令",[38,5167,5168],{},"节点参数可视化配置，temperature \u002F chunk size 等滑块调整直观",[38,5170,5171],{},"社区活跃，模板市场有不少现成的 Chatflow 可复用",[25,5173,5174],{},[41,5175,195],{},[35,5177,5178,5181,5184,5187,5190,5193,5196],{},[38,5179,5180],{},"流程复杂后画布混乱，连线交叉难以维护",[38,5182,5183],{},"性能一般——每个请求经过多个节点序列化\u002F反序列化，延迟偏高",[38,5185,5186],{},"错误信息不够友好，节点报错时定位问题费时",[38,5188,5189],{},"升级 LangChain 版本后部分节点可能不兼容",[38,5191,5192],{},"无法版本控制流程结构，团队协作时容易覆盖",[38,5194,5195],{},"并发能力有限，高并发场景需配合队列 + 负载均衡",[38,5197,5198],{},"深度定制仍需写代码——自定义节点门槛不低",[20,5200,221],{"id":221},[223,5202,5203,5207,5213,5216,5219,5222],{},[38,5204,800,5205,804],{},[228,5206,803],{},[38,5208,5209,5210,811],{},"访问 ",[228,5211,5212],{},"http:\u002F\u002Flocalhost:3000",[38,5214,5215],{},"配置 Credential：添加 OpenAI API Key 等凭据",[38,5217,5218],{},"新建 Chatflow → 从空白或模板开始",[38,5220,5221],{},"拖入节点：Chat OpenAI + Conversational Retrieval Chain + Vector Store",[38,5223,5224],{},"连线编排 → 右上角 Save → 测试对话 → 导出 API \u002F 嵌入组件",[20,5226,266],{"id":266},[97,5228,5229,5243],{},[100,5230,5231],{},[103,5232,5233,5235,5237,5239,5241],{},[106,5234,275],{},[106,5236,531],{},[106,5238,843],{},[106,5240,286],{},[106,5242,2321],{},[115,5244,5245,5262,5278,5290,5304,5316,5329,5343],{},[103,5246,5247,5250,5253,5256,5259],{},[120,5248,5249],{},"编排方式",[120,5251,5252],{},"拖拽 Chatflow",[120,5254,5255],{},"拖拽 Flow",[120,5257,5258],{},"拖拽 + YAML",[120,5260,5261],{},"拖拽 Workflow",[103,5263,5264,5267,5270,5272,5275],{},[120,5265,5266],{},"生态",[120,5268,5269],{},"LangChain",[120,5271,5269],{},[120,5273,5274],{},"自有",[120,5276,5277],{},"通用自动化",[103,5279,5280,5282,5284,5286,5288],{},[120,5281,221],{},[120,5283,296],{},[120,5285,296],{},[120,5287,296],{},[120,5289,855],{},[103,5291,5292,5294,5297,5299,5302],{},[120,5293,2332],{},[120,5295,5296],{},"✅ 模板丰富",[120,5298,370],{},[120,5300,5301],{},"✅ 强",[120,5303,367],{},[103,5305,5306,5308,5310,5312,5314],{},[120,5307,2971],{},[120,5309,370],{},[120,5311,370],{},[120,5313,5301],{},[120,5315,370],{},[103,5317,5318,5321,5323,5325,5327],{},[120,5319,5320],{},"API 导出",[120,5322,370],{},[120,5324,370],{},[120,5326,370],{},[120,5328,370],{},[103,5330,5331,5334,5337,5339,5341],{},[120,5332,5333],{},"部署",[120,5335,5336],{},"Docker",[120,5338,5336],{},[120,5340,5336],{},[120,5342,5336],{},[103,5344,5345,5348,5351,5353,5356],{},[120,5346,5347],{},"适合",[120,5349,5350],{},"AI 应用原型",[120,5352,5350],{},[120,5354,5355],{},"生产 AI 应用",[120,5357,5277],{},[20,5359,405],{"id":405},[35,5361,5362,5368,5374,5380,5386,5392,5398],{},[38,5363,5364,5367],{},[41,5365,5366],{},"流程别太复杂","：超过 15 个节点的 Chatflow 维护成本急升，拆分成多个",[38,5369,5370,5373],{},[41,5371,5372],{},"性能优化","：合并可合并的节点，减少序列化开销",[38,5375,5376,5379],{},[41,5377,5378],{},"版本管理","：定期导出 Chatflow JSON 备份，升级前测兼容性",[38,5381,5382,5385],{},[41,5383,5384],{},"凭据安全","：不要在流程中硬编码 API Key，统一走 Credential 管理",[38,5387,5388,5391],{},[41,5389,5390],{},"并发测试","：上线前压测，Flowise 单实例并发有限",[38,5393,5394,5397],{},[41,5395,5396],{},"LangChain 版本","：关注 Flowise 更新日志，LangChain 大版本升级可能有 breaking change",[38,5399,5400,5403],{},[41,5401,5402],{},"别替代生产框架","：复杂生产应用还是用 Dify 或直接写代码",[20,5405,457],{"id":456},[35,5407,5408,5411,5414,5417,5420,5423,5426,5429,5432,5435],{},[38,5409,5410],{},"✅ 不写代码搭建 RAG \u002F Chatbot 原型",[38,5412,5413],{},"✅ LangChain 生态用户可视化探索",[38,5415,5416],{},"✅ 快速验证 AI 应用概念",[38,5418,5419],{},"✅ 嵌入式聊天组件场景",[38,5421,5422],{},"✅ 教学 \u002F 演示 LLM 流程",[38,5424,5425],{},"❌ 复杂业务逻辑编排（用 n8n \u002F Dify）",[38,5427,5428],{},"❌ 大规模并发生产服务（性能有限）",[38,5430,5431],{},"❌ 深度定制需求（直接写 LangChain 代码）",[38,5433,5434],{},"❌ 需要版本控制 + 团队协作开发（能力有限）",[38,5436,5437],{},"❌ 非 LangChain 生态需求",[20,5439,493],{"id":492},[25,5441,5442,5445],{},[41,5443,5444],{},"Q: Flowise 和 Langflow 怎么选？","\nA: 两者定位几乎相同——都是 LangChain 可视化编排。Flowise 界面更简洁、上手稍快、社区模板多。Langflow 由 DataStax 维护、与 LangChain 官方关系更近、组件更新更快。都试试选顺手的即可。",[25,5447,5448,5451],{},[41,5449,5450],{},"Q: Flowise 和 Dify 怎么选？","\nA: Flowise 专注 LLM 流程可视化编排，轻量、原型验证快。Dify 是完整 AI 应用开发平台——工作流 + Agent + RAG + API 管理 + 监控，功能更全更适合生产。做原型选 Flowise，做产品选 Dify。",[25,5453,5454,5456,5457,5460],{},[41,5455,510],{},"\nA: 可以。Flowise 支持 Ollama \u002F HuggingFace 本地模型节点，配置 Ollama 地址（",[228,5458,5459],{},"http:\u002F\u002Flocalhost:11434","）即可在流程中使用本地 LLM 和 Embedding 模型。",[25,5462,5463,5466],{},[41,5464,5465],{},"Q: 生产环境能用吗？","\nA: 小规模可以（内部工具 \u002F 低并发场景）。高并发生产环境建议用 Dify 或直接写代码——Flowise 的节点序列化开销和单实例并发限制是瓶颈。如需生产部署，配合 Nginx 负载均衡 + 多实例 + Redis 队列。",[20,5468,520],{"id":520},[25,5470,5471,527,5473,527,5475],{},[524,5472,280],{"href":526},[524,5474,10],{"href":1532},[524,5476,1053],{"href":1052},[20,5478,538],{"id":538},[151,5480,5481],{},[25,5482,543],{},[35,5484,5485,5491],{},[38,5486,5487],{},[524,5488,553],{"href":5489,"rel":5490},"https:\u002F\u002Fflowiseai.com",[552],[38,5492,5493],{},[524,5494,560],{"href":5495,"rel":5496},"https:\u002F\u002Fgithub.com\u002FFlowiseAI\u002FFlowise",[552],{"title":562,"searchDepth":563,"depth":563,"links":5498},[5499,5500,5501,5502,5503,5504,5505,5506,5507,5508,5509],{"id":22,"depth":566,"text":23},{"id":33,"depth":566,"text":33},{"id":95,"depth":566,"text":95},{"id":158,"depth":566,"text":159},{"id":221,"depth":566,"text":221},{"id":266,"depth":566,"text":266},{"id":405,"depth":566,"text":405},{"id":456,"depth":566,"text":457},{"id":492,"depth":566,"text":493},{"id":520,"depth":566,"text":520},{"id":538,"depth":566,"text":538},"\u002Fimg\u002Ftools\u002Fflowise.webp","Flowise 真实评测：开源 LLM 流程编排平台（Apache 2.0 协议），拖拽式可视化构建 AI 应用，基于 LangChain 生态。支持 Docker 自托管 + Cloud 云端，适合不写代码也能搭建 RAG\u002FAgent\u002FChatbot 流程的团队。",[583],{},"\u002Ftools\u002Fagent\u002Fplatform\u002Fflowise",[590,591],"Free \u002F 开源（Apache 2.0）\u002F Cloud",{"power":563,"ux":595,"price":596,"cn_support":563,"stability":563},{"title":531,"description":5511},"Flowise - 拖拽式 LLM 流程编排评测 | AIHO","agent\u002Fplatform\u002Fflowise",[5522,5523],{"title":553,"url":5489},{"title":560,"url":5495},"tools\u002Fagent\u002Fplatform\u002Fflowise","开源 LLM 流程编排，拖拽式构建 AI 应用",[606,610,5527,2497,5528],"flow","langchain","不写代码也要拖拽搭建 RAG \u002F Chatbot \u002F Agent 流程的团队首选，基于 LangChain 生态组件丰富、可视化编排直观，但复杂流程维护难、性能一般、深度定制仍需写代码。","T4ETS7kNZnBNcywZPndJ_ynLVGjoghi1D0jMxZ7sK1A",{"id":5532,"title":843,"alternatives":5533,"api_compatible":15,"body":5536,"category":577,"chinese_friendly":563,"cover":6094,"description":6095,"domestic":580,"extension":581,"faq":6096,"free":580,"github":15,"languages":6109,"lastVerified":15,"meta":6111,"models":15,"navigation":586,"notSuitable":15,"opensource":586,"path":6112,"pillar":588,"platforms":6113,"priceTable":6115,"pricing":6128,"published":6129,"relatedPlaybooks":6130,"relatedReviews":15,"score":6132,"self_host":586,"seo":6133,"seoTitle":6134,"slug":12,"sources":6135,"stem":6145,"suitable":15,"tagline":6146,"tags":6147,"updated":2498,"verdict":6150,"website":6138,"__hash__":6151},"tools\u002Ftools\u002Fagent\u002Fplatform\u002Flangflow.md",[14,5534,5535],"agent\u002Fprotocol\u002Fcomposio","agent\u002Fgeneral\u002Fopenmanus",{"type":17,"value":5537,"toc":6082},[5538,5540,5543,5546,5548,5622,5624,5649,5654,5658,5662,5688,5692,5718,5720,5776,5779,5802,5804,5941,5943,5998,6000,6026,6028,6048,6050,6080],[20,5539,23],{"id":22},[25,5541,5542],{},"Langflow 是 2023 开源、2024 被 DataStax 收购的可视化 LangChain 画布。20,000+ GitHub stars、MIT 协议、Python 实现、与 Astra DB 向量库深度集成。差异点：拖拽式画布把 LangChain primitive 映射成节点 + RAG pipeline 原生组件 + 多 agent 工作流 + 节点可下钻到 Python 代码 + 自托管 \u002F Docker \u002F DataStax Astra 云托管 + Pinecone \u002F pgvector \u002F 主流向量库适配 + Visual GUI for building LangChain pipelines。Self-host 免费 \u002F Cloud Free + Paid ~$25\u002F月起。",[25,5544,5545],{},"适合：工程团队要可视化建 LangChain 流 + 不被 SaaS 锁住；合规 \u002F 数据驻留要求自托管；要 Astra DB 一站式 RAG；想用 LangChain 但讨厌纯代码协作。不适合：业务侧 + 非工程师（用 Dify \u002F Coze）；纯无代码偏好（Langflow 仍要懂 LangChain 概念）；轻量场景 + 不需要画布（直接写 LangChain 代码）。",[20,5547,33],{"id":33},[35,5549,5550,5556,5562,5568,5574,5580,5586,5592,5598,5604,5610,5616],{},[38,5551,5552,5555],{},[41,5553,5554],{},"可视化 LangChain 画布","：节点 = primitive，连线 = 数据流",[38,5557,5558,5561],{},[41,5559,5560],{},"RAG pipeline 原生组件","：Chunking \u002F Embedding \u002F VectorStore \u002F Retriever",[38,5563,5564,5567],{},[41,5565,5566],{},"多 agent 工作流","：编排多 agent 协作",[38,5569,5570,5573],{},[41,5571,5572],{},"Python 下钻","：任意节点可写 custom Python",[38,5575,5576,5579],{},[41,5577,5578],{},"向量库适配","：Astra DB \u002F Pinecone \u002F pgvector \u002F Weaviate \u002F Chroma",[38,5581,5582,5585],{},[41,5583,5584],{},"API 部署","：流程一键导出为 REST API",[38,5587,5588,5591],{},[41,5589,5590],{},"Real-time collaboration","：多用户同 project",[38,5593,5594,5597],{},[41,5595,5596],{},"版本控制","：内置 versioning + revert",[38,5599,5600,5603],{},[41,5601,5602],{},"数据可视化","：node output \u002F data flow 可视化调试",[38,5605,5606,5609],{},[41,5607,5608],{},"角色权限","：user auth + RBAC",[38,5611,5612,5615],{},[41,5613,5614],{},"Docker \u002F pip 安装","：5 分钟启动",[38,5617,5618,5621],{},[41,5619,5620],{},"Astra-hosted cloud","：DataStax 托管选项",[20,5623,95],{"id":95},[35,5625,5626,5632,5638,5644],{},[38,5627,5628,5631],{},[41,5629,5630],{},"Self-Host (OSS)","：$0；MIT 完全免费，自付 LLM API + 服务器",[38,5633,5634,5637],{},[41,5635,5636],{},"Cloud Free","：$0；DataStax Astra-hosted 小流量",[38,5639,5640,5643],{},[41,5641,5642],{},"Cloud Paid","：~$25\u002F月起；Astra DB + 更高额度 + 团队协作",[38,5645,5646,5648],{},[41,5647,728],{},"：联系销售；SSO + audit + 私有部署 + SLA",[151,5650,5651],{},[25,5652,5653],{},"自托管最低成本：$5-10\u002F月 VPS + LLM API token。Cloud Paid $25\u002F月适合不想运维的小团队。",[20,5655,5657],{"id":5656},"实测rag-应用-内部工具搭建","实测（RAG 应用 + 内部工具搭建）",[25,5659,5660],{},[41,5661,167],{},[35,5663,5664,5667,5670,5673,5676,5679,5682,5685],{},[38,5665,5666],{},"画布直观，比纯写 LangChain 协作效率高 5x",[38,5668,5669],{},"节点下钻到 Python 让灵活度不被画布限制",[38,5671,5672],{},"Astra DB 集成省了配 vector store 时间",[38,5674,5675],{},"4.4\u002F5 用户评分（Propicked \u002F Tooliverse）",[38,5677,5678],{},"开源 + 自托管 + 数据驻留满足合规",[38,5680,5681],{},"多 agent 编排比裸 LangChain 调试容易",[38,5683,5684],{},"RAG pipeline 模板一键起 demo",[38,5686,5687],{},"与 DataStax 长期支持降低 abandon ware 风险",[25,5689,5690],{},[41,5691,195],{},[35,5693,5694,5697,5700,5703,5706,5709,5712,5715],{},[38,5695,5696],{},"稳定性偶发：复杂流大流量下 node 偶尔失联（SelectHub 反馈）",[38,5698,5699],{},"第三方 API 依赖：external API 失败时错误处理弱",[38,5701,5702],{},"production readiness 不算 mission-critical（要自加 observability）",[38,5704,5705],{},"LangChain 升级偶尔 break 旧 flow",[38,5707,5708],{},"文档对新组件滞后 1-2 月",[38,5710,5711],{},"中文 UI 不完整，业务侧用户上手陡",[38,5713,5714],{},"大型 flow（100+ 节点）画布卡顿",[38,5716,5717],{},"多人协作偶发同步冲突",[20,5719,221],{"id":221},[3185,5721,5723],{"className":3187,"code":5722,"language":3189,"meta":562,"style":562},"# pip 安装\npip install langflow\nlangflow run  # http:\u002F\u002Flocalhost:7860\n\n# 或 Docker\ndocker run -p 7860:7860 langflowai\u002Flangflow:latest\n",[228,5724,5725,5730,5741,5752,5756,5761],{"__ignoreMap":562},[3193,5726,5727],{"class":3195,"line":3196},[3193,5728,5729],{"class":3245},"# pip 安装\n",[3193,5731,5732,5735,5738],{"class":3195,"line":566},[3193,5733,5734],{"class":3199},"pip",[3193,5736,5737],{"class":3203}," install",[3193,5739,5740],{"class":3203}," langflow\n",[3193,5742,5743,5746,5749],{"class":3195,"line":563},[3193,5744,5745],{"class":3199},"langflow",[3193,5747,5748],{"class":3203}," run",[3193,5750,5751],{"class":3245},"  # http:\u002F\u002Flocalhost:7860\n",[3193,5753,5754],{"class":3195,"line":595},[3193,5755,3391],{"emptyLinePlaceholder":586},[3193,5757,5758],{"class":3195,"line":596},[3193,5759,5760],{"class":3245},"# 或 Docker\n",[3193,5762,5763,5765,5767,5770,5773],{"class":3195,"line":3388},[3193,5764,591],{"class":3199},[3193,5766,5748],{"class":3203},[3193,5768,5769],{"class":3212}," -p",[3193,5771,5772],{"class":3203}," 7860:7860",[3193,5774,5775],{"class":3203}," langflowai\u002Flangflow:latest\n",[25,5777,5778],{},"试 RAG 流：",[223,5780,5781,5784,5787,5790,5793,5796,5799],{},[38,5782,5783],{},"新建 flow → 选 Document QA 模板",[38,5785,5786],{},"Document Loader 节点 → 上传 PDF",[38,5788,5789],{},"Splitter → Embedder（OpenAI 或本地）",[38,5791,5792],{},"VectorStore（Astra \u002F Chroma）",[38,5794,5795],{},"Retriever + ChatOpenAI → Chat Output",[38,5797,5798],{},"部署为 API → 拿到 endpoint",[38,5800,5801],{},"复杂场景下钻节点写 Python 自定义",[20,5803,266],{"id":266},[97,5805,5806,5820],{},[100,5807,5808],{},[103,5809,5810,5812,5814,5816,5818],{},[106,5811,275],{},[106,5813,843],{},[106,5815,286],{},[106,5817,2321],{},[106,5819,531],{},[115,5821,5822,5839,5853,5867,5880,5896,5909,5926],{},[103,5823,5824,5827,5830,5833,5836],{},[120,5825,5826],{},"中心",[120,5828,5829],{},"LangChain primitive",[120,5831,5832],{},"LLMOps 全平台",[120,5834,5835],{},"通用 workflow",[120,5837,5838],{},"LangChain（JS）",[103,5840,5841,5843,5845,5848,5851],{},[120,5842,2018],{},[120,5844,4580],{},[120,5846,5847],{},"✅ AGPL",[120,5849,5850],{},"✅ Sustainable",[120,5852,4580],{},[103,5854,5855,5857,5860,5863,5865],{},[120,5856,3291],{},[120,5858,5859],{},"✅ pip\u002FDocker",[120,5861,5862],{},"✅ Docker",[120,5864,5862],{},[120,5866,370],{},[103,5868,5869,5871,5874,5876,5878],{},[120,5870,321],{},[120,5872,5873],{},"✅ 旗舰",[120,5875,370],{},[120,5877,370],{},[120,5879,370],{},[103,5881,5882,5885,5888,5891,5894],{},[120,5883,5884],{},"代码下钻",[120,5886,5887],{},"✅ Python",[120,5889,5890],{},"部分",[120,5892,5893],{},"✅ JS",[120,5895,5893],{},[103,5897,5898,5901,5903,5905,5907],{},[120,5899,5900],{},"RAG 内置",[120,5902,370],{},[120,5904,370],{},[120,5906,5890],{},[120,5908,370],{},[103,5910,5911,5914,5917,5920,5923],{},[120,5912,5913],{},"起价（云）",[120,5915,5916],{},"$25\u002F月",[120,5918,5919],{},"$59\u002F月（Team）",[120,5921,5922],{},"自托管 $0",[120,5924,5925],{},"–",[103,5927,5928,5930,5933,5936,5938],{},[120,5929,5347],{},[120,5931,5932],{},"工程 + LangChain",[120,5934,5935],{},"业务 + LLMOps",[120,5937,5277],{},[120,5939,5940],{},"JS 生态",[20,5942,405],{"id":405},[35,5944,5945,5951,5957,5963,5969,5975,5981,5987,5992],{},[38,5946,5947,5950],{},[41,5948,5949],{},"自托管推荐 Docker","：pip 版本依赖冲突难调",[38,5952,5953,5956],{},[41,5954,5955],{},"生产加 observability","：langsmith \u002F sentry \u002F prometheus 必装",[38,5958,5959,5962],{},[41,5960,5961],{},"Astra DB cloud free","：起步够用，付费版起步前算清成本",[38,5964,5965,5968],{},[41,5966,5967],{},"复杂 flow 拆模块","：100+ 节点画布卡顿，拆成子 flow",[38,5970,5971,5974],{},[41,5972,5973],{},"LangChain 版本 pin","：Langflow 升级前测试 flow 兼容性",[38,5976,5977,5980],{},[41,5978,5979],{},"第三方 API 加重试","：custom Python 节点写 retry + fallback",[38,5982,5983,5986],{},[41,5984,5985],{},"多人协作 lock","：同时编辑 flow 容易冲突，加 lock 流程",[38,5988,5989,5991],{},[41,5990,3577],{},"：UI 英文为主，业务侧用户先培训",[38,5993,5994,5997],{},[41,5995,5996],{},"Self-host vs Cloud","：合规要 self-host，省心要 Cloud",[20,5999,457],{"id":456},[35,6001,6002,6005,6008,6011,6014,6017,6020,6023],{},[38,6003,6004],{},"✅ 工程团队要可视化建 LangChain 流",[38,6006,6007],{},"✅ 合规 \u002F 数据驻留要求自托管",[38,6009,6010],{},"✅ 要 Astra DB 一站式 RAG",[38,6012,6013],{},"✅ Python 团队 + 想画布 + 想下钻代码",[38,6015,6016],{},"❌ 业务侧 + 非工程师（用 Dify \u002F Coze）",[38,6018,6019],{},"❌ 纯无代码偏好",[38,6021,6022],{},"❌ 轻量场景 + 直接写 LangChain 更快",[38,6024,6025],{},"❌ JS 生态优先（用 Flowise）",[20,6027,520],{"id":520},[35,6029,6030,6036,6042],{},[38,6031,6032],{},[524,6033,6035],{"href":6034},"\u002Ftools\u002Fagent\u002Fplatform\u002Fn8n","n8n 评测",[38,6037,6038],{},[524,6039,6041],{"href":6040},"\u002Ftools\u002Fagent\u002Fprotocol\u002Fcomposio","Composio 评测",[38,6043,6044],{},[524,6045,6047],{"href":6046},"\u002Ftools\u002Fagent\u002Fgeneral\u002Fopenmanus","OpenManus 评测",[20,6049,538],{"id":538},[223,6051,6052,6059,6066,6073],{},[38,6053,6054,6055],{},"Langflow 官网 + 定价 ",[524,6056,6057],{"href":6057,"rel":6058},"https:\u002F\u002Fwww.langflow.org\u002Fpricing",[552],[38,6060,6061,6062],{},"AutomationAtlas — Langflow $25\u002Fmo 2026 Comparison ",[524,6063,6064],{"href":6064,"rel":6065},"https:\u002F\u002Fautomationatlas.io\u002Ftools\u002Flangflow",[552],[38,6067,6068,6069],{},"xpay — LangFlow 2026 Visual AI Agent Builder（MIT + DataStax）",[524,6070,6071],{"href":6071,"rel":6072},"https:\u002F\u002Fwww.xpay.sh\u002Fresources\u002Fagentic-frameworks\u002Flangflow",[552],[38,6074,6075,6076],{},"SelectHub — LangFlow Reviews 2026（稳定性 + 生产 readiness）",[524,6077,6078],{"href":6078,"rel":6079},"https:\u002F\u002Fwww.selecthub.com\u002Fp\u002Fai-agent-builder-software\u002Flangflow\u002F",[552],[3864,6081,4924],{},{"title":562,"searchDepth":563,"depth":563,"links":6083},[6084,6085,6086,6087,6088,6089,6090,6091,6092,6093],{"id":22,"depth":566,"text":23},{"id":33,"depth":566,"text":33},{"id":95,"depth":566,"text":95},{"id":5656,"depth":566,"text":5657},{"id":221,"depth":566,"text":221},{"id":266,"depth":566,"text":266},{"id":405,"depth":566,"text":405},{"id":456,"depth":566,"text":457},{"id":520,"depth":566,"text":520},{"id":538,"depth":566,"text":538},"\u002Fimg\u002Ftools\u002Flangflow.webp","Langflow 真实评测：2023 开源项目，2024 被 DataStax 收购，与 Astra DB 向量数据库平台深度集成。20,000+ GitHub stars、MIT 协议。差异点：拖拽式可视化画布映射 LangChain 原语 + RAG pipeline 原生组件 + 多 agent 工作流 + Python + 自托管（pip \u002F Docker）+ DataStax Astra 托管云 + Pinecone \u002F pgvector \u002F 主流向量库适配。Cloud 免费档 + 约 $25\u002F月起。",[6097,6100,6103,6106],{"q":6098,"a":6099},"Langflow 和 Dify \u002F n8n \u002F Flowise 怎么选？","Langflow 强『可视化 LangChain 原语 + Python 代码可下钻 + 自托管 + Astra DB 集成』，工程导向。Dify 是 LLMOps 全平台（含 dataset \u002F app \u002F observability），业务侧更友好。n8n 是通用 workflow（非 LangChain 中心），70+ LangChain 节点是后加的。Flowise 是 Langflow 的同类竞品（JS 生态）。工程团队 + Python + 合规自托管 → Langflow；业务 + 完整 LLMOps → Dify；通用自动化 → n8n。",{"q":6101,"a":6102},"为什么被 DataStax 收购？","DataStax 是 Cassandra 商业公司 + Astra DB 向量数据库厂商。收购 Langflow 是为了把『可视化 LangChain builder』和『生产级 vector store』捆成一站式 RAG 解决方案。Langflow 主仓仍是 MIT 开源，但 cloud \u002F 企业版深度依赖 Astra。",{"q":6104,"a":6105},"可视化画布会不会限制灵活度？","Langflow 节点对应 LangChain primitive，可在任意节点下钻到 Python 代码 + 自定义。所以可视化层是『脚手架 + 协作工具』，不是『纯无代码黑盒』。复杂逻辑、custom tool、retriever 都能写代码扩展。",{"q":6107,"a":6108},"production readiness 如何？","用户反馈：原型 + 内部工具非常顺；大流量 \u002F 关键业务要自行加 observability \u002F 错误处理 \u002F 缓存。SelectHub 评测列出『稳定性偶发 + 第三方 API 依赖 + 非完全 production-ready』。生产部署建议 Astra-hosted cloud 或自托管 + 加 sentry \u002F langsmith \u002F prometheus。",[583,6110],"multi",{},"\u002Ftools\u002Fagent\u002Fplatform\u002Flangflow",[1112,6114,591,2440],"cloud",[6116,6119,6121,6125],{"plan":5630,"price":125,"features":6117,"notes":6118},"MIT + pip\u002FDocker + 全部组件 + Pinecone\u002Fpgvector\u002FAstra 适配","自付 LLM API + 服务器",{"plan":5636,"price":125,"features":6120,"notes":4966},"DataStax 托管 + 小流量",{"plan":5642,"price":6122,"features":6123,"notes":6124},"~$25\u002F月起","Astra DB + 更高额度 + 更多并发 + 团队协作","按使用量阶梯",{"plan":728,"price":146,"features":6126,"notes":6127},"SSO + audit + 私有部署 + SLA + 数据驻留","合规 \u002F 大客户","Self-host 完全免费 MIT \u002F Cloud Free + 付费起 ~$25·月（Astra-hosted）","2026-06-19",[6131],"onboarding\u002Frag-app-workflow",{"power":595,"ux":596,"price":596,"cn_support":563,"stability":595},{"title":843,"description":6095},"Langflow 评测 2026：可视化 AI 工作流构建工具，LangChain 低代码平台",[6136,6139,6141,6143],{"name":6137,"url":6138,"accessed":2498},"Langflow 官网","https:\u002F\u002Fwww.langflow.org",{"name":6140,"url":6064,"accessed":2498},"AutomationAtlas — Langflow $25\u002Fmo 2026 Comparison",{"name":6142,"url":6071,"accessed":2498},"xpay — LangFlow 2026 Visual AI Agent Builder",{"name":6144,"url":6078,"accessed":2498},"SelectHub — LangFlow Reviews 2026","tools\u002Fagent\u002Fplatform\u002Flangflow","DataStax 出品的可视化 LangChain 画布——MIT 开源 + 20k+ stars + 自托管 + Astra DB 云托管",[610,6148,5528,1113,6149,5745],"visual-builder","datastax","工程团队要可视化建 LangChain 流 + 不被 SaaS 锁住 + 合规要求自托管的最佳选择。要纯无代码 + 业务侧 → Dify \u002F Coze；要纯代码 → 直接写 LangChain \u002F LlamaIndex。","rvh-hO12QKzN5KXHjH_xWXuls1hBO42dYKXttkerPls",{"id":6153,"title":2321,"alternatives":6154,"api_compatible":15,"body":6155,"category":577,"chinese_friendly":563,"cover":6801,"description":6802,"domestic":580,"extension":581,"faq":6803,"free":580,"github":15,"languages":6816,"lastVerified":15,"meta":6817,"models":15,"navigation":586,"notSuitable":15,"opensource":586,"path":6034,"pillar":588,"platforms":6818,"priceTable":6819,"pricing":6835,"published":6129,"relatedPlaybooks":6836,"relatedReviews":15,"score":6838,"self_host":586,"seo":6839,"seoTitle":6840,"slug":14,"sources":6841,"stem":6850,"suitable":15,"tagline":6851,"tags":6852,"updated":2498,"verdict":6854,"website":6763,"__hash__":6855},"tools\u002Ftools\u002Fagent\u002Fplatform\u002Fn8n.md",[12,5534,5535],{"type":17,"value":6156,"toc":6789},[6157,6159,6162,6165,6167,6241,6243,6266,6271,6275,6279,6305,6309,6335,6339,6464,6467,6495,6498,6500,6646,6648,6709,6711,6737,6739,6754,6756,6786],[20,6158,23],{"id":22},[25,6160,6161],{},"n8n 是 2019 创立、Sustainable Use License（fair-code）的开源自动化平台。2026 突破 200,000 active users、5x ARR 增长、5,800+ 社区 AI workflow。差异点：近 70 个 LangChain 专属节点 + 原生 MCP 协议 + AI Agent 节点（reasoning loop + 工具调用）+ Ollama \u002F OpenAI 双路 + 400+ 集成 + execution-based 定价（步骤数无关）+ 自托管 VPS $5-10\u002F月跑全套。",[25,6163,6164],{},"适合：开发者 + 想完全控制 + 高量级自动化；从 Zapier \u002F Make 迁出降本；要 AI Agent + LangChain + MCP 一体；合规 \u002F 自托管 \u002F 数据驻留要求。不适合：非技术 + 要 8000+ 现成集成（用 Zapier）；中等复杂 + 不愿自托管（用 Make）；纯研究 \u002F 学术 agent（用 OpenManus \u002F Langflow）。",[20,6166,33],{"id":33},[35,6168,6169,6175,6181,6187,6193,6199,6205,6211,6217,6223,6229,6235],{},[38,6170,6171,6174],{},[41,6172,6173],{},"AI Agent 节点","：reasoning loop + 自主工具选择",[38,6176,6177,6180],{},[41,6178,6179],{},"70 LangChain 节点","：LLM \u002F VectorStore \u002F Agent \u002F Tool \u002F Memory 全套",[38,6182,6183,6186],{},[41,6184,6185],{},"原生 MCP","：MCP server 一键挂载到 agent",[38,6188,6189,6192],{},[41,6190,6191],{},"Ollama 集成","：本地 LLM 零 API 成本",[38,6194,6195,6198],{},[41,6196,6197],{},"400+ 集成","：Slack \u002F GitHub \u002F Google \u002F Notion \u002F 主流 SaaS",[38,6200,6201,6204],{},[41,6202,6203],{},"HTTP \u002F Webhook 万能节点","：任意 REST API 都能接",[38,6206,6207,6210],{},[41,6208,6209],{},"Cron \u002F Trigger","：定时 \u002F 事件 \u002F Webhook 触发",[38,6212,6213,6216],{},[41,6214,6215],{},"多分支并行 + 错误处理","：production workflow 必备",[38,6218,6219,6222],{},[41,6220,6221],{},"版本控制 + Git Sync","：workflow as code",[38,6224,6225,6228],{},[41,6226,6227],{},"自托管 Docker \u002F Kubernetes","：一行起 + 水平扩展",[38,6230,6231,6234],{},[41,6232,6233],{},"execution-based 定价","：20 步和 2 步同价（自托管 = 0）",[38,6236,6237,6240],{},[41,6238,6239],{},"5800+ 社区 workflow","：clone 即用",[20,6242,95],{"id":95},[35,6244,6245,6251,6256,6261],{},[38,6246,6247,6250],{},[41,6248,6249],{},"Community Self-host","：$0；全功能 + 不限 execution + 自付 VPS $5-10\u002F月",[38,6252,6253,6255],{},[41,6254,5110],{},"：~€20\u002F月；2,500 executions + 5 workflow",[38,6257,6258,6260],{},[41,6259,5121],{},"：~€50\u002F月；高 executions + 团队协作",[38,6262,6263,6265],{},[41,6264,728],{},"：联系销售；SSO + LDAP + 私有部署 + SLA",[151,6267,6268],{},[25,6269,6270],{},"真实场景：Zapier $50\u002F月跑中等复杂 → n8n 自托管 $5\u002F月跑同样的 = 10x 降本。",[20,6272,6274],{"id":6273},"实测中型团队-saas-迁移-本地-ai","实测（中型团队 SaaS 迁移 + 本地 AI）",[25,6276,6277],{},[41,6278,167],{},[35,6280,6281,6284,6287,6290,6293,6296,6299,6302],{},[38,6282,6283],{},"自托管成本几乎可忽略：$5\u002F月 VPS 跑几十个 workflow",[38,6285,6286],{},"AI Agent + Ollama 让 LLM 任务零 API 成本",[38,6288,6289],{},"70 LangChain 节点覆盖 RAG \u002F Agent \u002F 多模态",[38,6291,6292],{},"MCP 原生集成让 n8n agent 调用任意 MCP server",[38,6294,6295],{},"5800+ 社区 workflow 节省 80% 上手时间",[38,6297,6298],{},"HTTP \u002F Webhook 万能节点弥补 native 集成缺口",[38,6300,6301],{},"升级 Docker tag 一行，无 vendor 升级费",[38,6303,6304],{},"中文社区 \u002F B 站教程丰富",[25,6306,6307],{},[41,6308,195],{},[35,6310,6311,6314,6317,6320,6323,6326,6329,6332],{},[38,6312,6313],{},"集成数（400+）远少于 Zapier（8000+），冷门 SaaS 要写 HTTP 自己接",[38,6315,6316],{},"自托管要懂 Docker \u002F Postgres \u002F Redis（高吞吐场景）",[38,6318,6319],{},"Cloud 定价 execution 计算法与本地不一致，迁移要重算成本",[38,6321,6322],{},"复杂 workflow 调试比 Zapier 难（错误堆栈深）",[38,6324,6325],{},"Sustainable Use License 不是传统 OSI 开源，企业法务要看条款",[38,6327,6328],{},"AI Agent 节点 production 稳定性不如简单线性 flow",[38,6330,6331],{},"Webhook 公网暴露要加 IP 白名单 + secret",[38,6333,6334],{},"Worker 模式才能并发，单进程吞吐有限",[20,6336,6338],{"id":6337},"上手docker-5-分钟","上手（Docker 5 分钟）",[3185,6340,6342],{"className":3187,"code":6341,"language":3189,"meta":562,"style":562},"# 持久化目录\nmkdir -p ~\u002Fn8n-data\n\n# 启动\ndocker run -d \\\n  --name n8n \\\n  -p 5678:5678 \\\n  -e N8N_BASIC_AUTH_ACTIVE=true \\\n  -e N8N_BASIC_AUTH_USER=admin \\\n  -e N8N_BASIC_AUTH_PASSWORD=yourpassword \\\n  -v ~\u002Fn8n-data:\u002Fhome\u002Fnode\u002F.n8n \\\n  --restart always \\\n  n8nio\u002Fn8n\n\n# http:\u002F\u002Flocalhost:5678\n",[228,6343,6344,6349,6359,6363,6367,6379,6389,6399,6412,6421,6430,6440,6450,6455,6459],{"__ignoreMap":562},[3193,6345,6346],{"class":3195,"line":3196},[3193,6347,6348],{"class":3245},"# 持久化目录\n",[3193,6350,6351,6354,6356],{"class":3195,"line":566},[3193,6352,6353],{"class":3199},"mkdir",[3193,6355,5769],{"class":3212},[3193,6357,6358],{"class":3203}," ~\u002Fn8n-data\n",[3193,6360,6361],{"class":3195,"line":563},[3193,6362,3391],{"emptyLinePlaceholder":586},[3193,6364,6365],{"class":3195,"line":595},[3193,6366,4339],{"class":3245},[3193,6368,6369,6371,6373,6376],{"class":3195,"line":596},[3193,6370,591],{"class":3199},[3193,6372,5748],{"class":3203},[3193,6374,6375],{"class":3212}," -d",[3193,6377,6378],{"class":3212}," \\\n",[3193,6380,6381,6384,6387],{"class":3195,"line":3388},[3193,6382,6383],{"class":3212},"  --name",[3193,6385,6386],{"class":3203}," n8n",[3193,6388,6378],{"class":3212},[3193,6390,6391,6394,6397],{"class":3195,"line":3394},[3193,6392,6393],{"class":3212},"  -p",[3193,6395,6396],{"class":3203}," 5678:5678",[3193,6398,6378],{"class":3212},[3193,6400,6401,6404,6407,6410],{"class":3195,"line":3400},[3193,6402,6403],{"class":3212},"  -e",[3193,6405,6406],{"class":3203}," N8N_BASIC_AUTH_ACTIVE=",[3193,6408,6409],{"class":3212},"true",[3193,6411,6378],{"class":3212},[3193,6413,6414,6416,6419],{"class":3195,"line":3406},[3193,6415,6403],{"class":3212},[3193,6417,6418],{"class":3203}," N8N_BASIC_AUTH_USER=admin",[3193,6420,6378],{"class":3212},[3193,6422,6423,6425,6428],{"class":3195,"line":3411},[3193,6424,6403],{"class":3212},[3193,6426,6427],{"class":3203}," N8N_BASIC_AUTH_PASSWORD=yourpassword",[3193,6429,6378],{"class":3212},[3193,6431,6432,6435,6438],{"class":3195,"line":3417},[3193,6433,6434],{"class":3212},"  -v",[3193,6436,6437],{"class":3203}," ~\u002Fn8n-data:\u002Fhome\u002Fnode\u002F.n8n",[3193,6439,6378],{"class":3212},[3193,6441,6442,6445,6448],{"class":3195,"line":3422},[3193,6443,6444],{"class":3212},"  --restart",[3193,6446,6447],{"class":3203}," always",[3193,6449,6378],{"class":3212},[3193,6451,6452],{"class":3195,"line":3428},[3193,6453,6454],{"class":3203},"  n8nio\u002Fn8n\n",[3193,6456,6457],{"class":3195,"line":3434},[3193,6458,3391],{"emptyLinePlaceholder":586},[3193,6460,6461],{"class":3195,"line":4361},[3193,6462,6463],{"class":3245},"# http:\u002F\u002Flocalhost:5678\n",[25,6465,6466],{},"连 Ollama：",[3185,6468,6470],{"className":3187,"code":6469,"language":3189,"meta":562,"style":562},"ollama serve\nollama pull llama3.2\n# n8n 中添加 Ollama credential：http:\u002F\u002Fhost.docker.internal:11434\n",[228,6471,6472,6480,6490],{"__ignoreMap":562},[3193,6473,6474,6477],{"class":3195,"line":3196},[3193,6475,6476],{"class":3199},"ollama",[3193,6478,6479],{"class":3203}," serve\n",[3193,6481,6482,6484,6487],{"class":3195,"line":566},[3193,6483,6476],{"class":3199},[3193,6485,6486],{"class":3203}," pull",[3193,6488,6489],{"class":3203}," llama3.2\n",[3193,6491,6492],{"class":3195,"line":563},[3193,6493,6494],{"class":3245},"# n8n 中添加 Ollama credential：http:\u002F\u002Fhost.docker.internal:11434\n",[25,6496,6497],{},"试 workflow：Webhook 触发 → AI Agent 节点（Ollama）→ Slack 通知。复制粘贴一个社区 workflow 30 分钟跑通完整 AI 自动化。",[20,6499,266],{"id":266},[97,6501,6502,6518],{},[100,6503,6504],{},[103,6505,6506,6508,6510,6513,6516],{},[106,6507,275],{},[106,6509,2321],{},[106,6511,6512],{},"Zapier",[106,6514,6515],{},"Make",[106,6517,843],{},[115,6519,6520,6533,6545,6562,6574,6588,6600,6614,6630],{},[103,6521,6522,6524,6527,6529,6531],{},[120,6523,2018],{},[120,6525,6526],{},"✅ fair-code",[120,6528,381],{},[120,6530,381],{},[120,6532,4580],{},[103,6534,6535,6537,6539,6541,6543],{},[120,6536,3291],{},[120,6538,5873],{},[120,6540,381],{},[120,6542,381],{},[120,6544,370],{},[103,6546,6547,6550,6553,6556,6559],{},[120,6548,6549],{},"集成数",[120,6551,6552],{},"400+",[120,6554,6555],{},"8000+",[120,6557,6558],{},"2000+",[120,6560,6561],{},"LangChain 原语",[103,6563,6564,6566,6568,6570,6572],{},[120,6565,2328],{},[120,6567,5873],{},[120,6569,5890],{},[120,6571,5890],{},[120,6573,370],{},[103,6575,6576,6579,6582,6584,6586],{},[120,6577,6578],{},"LangChain 节点",[120,6580,6581],{},"✅ 70 个",[120,6583,381],{},[120,6585,381],{},[120,6587,1849],{},[103,6589,6590,6592,6594,6596,6598],{},[120,6591,3799],{},[120,6593,1849],{},[120,6595,381],{},[120,6597,381],{},[120,6599,5890],{},[103,6601,6602,6605,6608,6610,6612],{},[120,6603,6604],{},"Local LLM",[120,6606,6607],{},"✅ Ollama",[120,6609,381],{},[120,6611,381],{},[120,6613,370],{},[103,6615,6616,6619,6622,6625,6628],{},[120,6617,6618],{},"起价",[120,6620,6621],{},"$0 自托管",[120,6623,6624],{},"$29.99\u002F月",[120,6626,6627],{},"$9\u002F月",[120,6629,6621],{},[103,6631,6632,6634,6637,6640,6643],{},[120,6633,5347],{},[120,6635,6636],{},"开发者 + 高量级",[120,6638,6639],{},"非技术 + 简单",[120,6641,6642],{},"中等复杂",[120,6644,6645],{},"LangChain 工程",[20,6647,405],{"id":405},[35,6649,6650,6656,6662,6668,6674,6680,6686,6692,6698,6703],{},[38,6651,6652,6655],{},[41,6653,6654],{},"自托管装 Postgres + Redis","：默认 SQLite 高吞吐崩",[38,6657,6658,6661],{},[41,6659,6660],{},"Worker 模式","：高并发要起 worker container 才能并行",[38,6663,6664,6667],{},[41,6665,6666],{},"Webhook 加防护","：公网 Webhook 加 IP 白名单 \u002F secret \u002F nginx",[38,6669,6670,6673],{},[41,6671,6672],{},"数据加密","：n8n encryption key 设强随机值，备份要带 key",[38,6675,6676,6679],{},[41,6677,6678],{},"Cloud vs Self-host 成本","：>2k execution\u002F月 自托管更省",[38,6681,6682,6685],{},[41,6683,6684],{},"集成缺失","：冷门 SaaS 用 HTTP Request + curl 等价",[38,6687,6688,6691],{},[41,6689,6690],{},"AI Agent 稳定性","：生产关键流先用线性节点，agent 留给探索任务",[38,6693,6694,6697],{},[41,6695,6696],{},"license 法务","：Sustainable Use License 给法务看一遍，企业内部用没问题",[38,6699,6700,6702],{},[41,6701,6239],{},"：导入前看作者 + star 数 + 不要直接生产用，要 review",[38,6704,6705,6708],{},[41,6706,6707],{},"monitoring","：生产部署加 prometheus + 错误告警",[20,6710,457],{"id":456},[35,6712,6713,6716,6719,6722,6725,6728,6731,6734],{},[38,6714,6715],{},"✅ 开发者 + 完全控制 + 高量级自动化",[38,6717,6718],{},"✅ 从 Zapier \u002F Make 迁出降本",[38,6720,6721],{},"✅ AI Agent + LangChain + MCP 一体",[38,6723,6724],{},"✅ 合规 \u002F 数据驻留 \u002F 自托管需求",[38,6726,6727],{},"❌ 非技术 + 要 8000+ 现成集成（用 Zapier）",[38,6729,6730],{},"❌ 完全不愿自托管 + 不想付 Cloud",[38,6732,6733],{},"❌ 纯研究 \u002F 学术 agent（用 OpenManus）",[38,6735,6736],{},"❌ 极简 2-step 自动化（Zapier 更快）",[20,6738,520],{"id":520},[35,6740,6741,6746,6750],{},[38,6742,6743],{},[524,6744,6745],{"href":6112},"Langflow 评测",[38,6747,6748],{},[524,6749,6041],{"href":6040},[38,6751,6752],{},[524,6753,6047],{"href":6046},[20,6755,538],{"id":538},[223,6757,6758,6765,6772,6779],{},[38,6759,6760,6761],{},"n8n 官网 ",[524,6762,6763],{"href":6763,"rel":6764},"https:\u002F\u002Fn8n.io",[552],[38,6766,6767,6768],{},"AutomationByExperts — n8n 2026 200k users 5x ARR ",[524,6769,6770],{"href":6770,"rel":6771},"https:\u002F\u002Fautomationbyexperts.com\u002Fblog\u002Fn8n-ai-workflow-automation-guide-2026",[552],[38,6773,6774,6775],{},"Tutorials Technology — n8n + AI on Linux 2026（Docker + Ollama）",[524,6776,6777],{"href":6777,"rel":6778},"https:\u002F\u002Ftutorials.technology\u002Ftutorials\u002Fn8n-ai-workflows-linux-2026.html",[552],[38,6780,6781,6782],{},"Northflank — n8n Self-host Architecture + Pricing 2026 ",[524,6783,6784],{"href":6784,"rel":6785},"https:\u002F\u002Fnorthflank.com\u002Fblog\u002Fhow-to-self-host-n8n-setup-architecture-and-pricing-guide",[552],[3864,6787,6788],{},"html pre.shiki code .sJ8bj, html code.shiki .sJ8bj{--shiki-default:#6A737D;--shiki-dark:#6A737D}html pre.shiki code .sScJk, html code.shiki .sScJk{--shiki-default:#6F42C1;--shiki-dark:#B392F0}html pre.shiki code .sj4cs, html code.shiki .sj4cs{--shiki-default:#005CC5;--shiki-dark:#79B8FF}html pre.shiki code .sZZnC, html code.shiki .sZZnC{--shiki-default:#032F62;--shiki-dark:#9ECBFF}html .default .shiki span {color: var(--shiki-default);background: var(--shiki-default-bg);font-style: var(--shiki-default-font-style);font-weight: var(--shiki-default-font-weight);text-decoration: var(--shiki-default-text-decoration);}html .shiki span {color: var(--shiki-default);background: var(--shiki-default-bg);font-style: var(--shiki-default-font-style);font-weight: var(--shiki-default-font-weight);text-decoration: var(--shiki-default-text-decoration);}html .dark .shiki span {color: var(--shiki-dark);background: var(--shiki-dark-bg);font-style: var(--shiki-dark-font-style);font-weight: var(--shiki-dark-font-weight);text-decoration: var(--shiki-dark-text-decoration);}html.dark .shiki span {color: var(--shiki-dark);background: var(--shiki-dark-bg);font-style: var(--shiki-dark-font-style);font-weight: var(--shiki-dark-font-weight);text-decoration: var(--shiki-dark-text-decoration);}",{"title":562,"searchDepth":563,"depth":563,"links":6790},[6791,6792,6793,6794,6795,6796,6797,6798,6799,6800],{"id":22,"depth":566,"text":23},{"id":33,"depth":566,"text":33},{"id":95,"depth":566,"text":95},{"id":6273,"depth":566,"text":6274},{"id":6337,"depth":566,"text":6338},{"id":266,"depth":566,"text":266},{"id":405,"depth":566,"text":405},{"id":456,"depth":566,"text":457},{"id":520,"depth":566,"text":520},{"id":538,"depth":566,"text":538},"\u002Fimg\u002Ftools\u002Fn8n.webp","n8n 2026 真实评测：开源自托管自动化平台和 AI Agent 工作流工具，支持 LangChain 节点、MCP、Ollama、OpenAI、Webhook、400+ 集成和 execution-based 定价。本文对比 Zapier、Make、Dify，整理自托管成本、适合场景和避坑建议。",[6804,6807,6810,6813],{"q":6805,"a":6806},"n8n 和 Zapier \u002F Make 怎么选？","Zapier 8000+ 集成 + 最易上手 + 非技术团队最爱，但 $29.99\u002F月才 750 tasks + 每步独立计费 = 量大成本爆炸。Make 视觉画布 + 并行分支 + 2000+ 集成 + 智能打包步骤（10k ops $29\u002F月），中等复杂最佳性价比。n8n 开源 + 自托管 + execution-based（步骤数无关）+ 70 LangChain 节点 + MCP 原生，开发者 + 高量级 + 完全控制首选。模式：从 Zapier \u002F Make 起步 → 撞墙 → 迁 n8n。",{"q":6808,"a":6809},"AI Agent 节点是什么？","n8n 2026 加的特殊节点：运行 reasoning loop，从连接的工具节点中自主挑选并调用，直到有答案。和传统线性节点的『按顺序执行预定动作』不同，AI Agent 引入了 LLM 决策。可挂接近 70 LangChain 节点 \u002F MCP server \u002F 任意 HTTP API。让 n8n 从『纯自动化』升级为『真正的 AI agent 平台』。",{"q":6811,"a":6812},"Sustainable Use License 是什么协议？","n8n 用的 fair-code 协议（非传统 OSI 开源）。允许内部使用 + 自托管 + 修改源码，但限制把 n8n 作为 SaaS 转售（与 n8n 商业版直接竞争）。对自用 \u002F 内部工具 \u002F 普通自托管 100% 免费。要做 n8n competitor \u002F 商业 SaaS 才需要谈授权。",{"q":6814,"a":6815},"自托管成本和门槛？","最小：$5-10\u002F月 VPS（DigitalOcean \u002F Hetzner \u002F Linode）+ Docker 一行起。Postgres 持久化 + Redis 队列（高吞吐）+ Worker 节点（水平扩展）。10 分钟内能跑通最小版本。中文社区 \u002F B 站 \u002F 知乎 有大量中文教程。比 Langflow \u002F Dify 上手快。",[583,6110],{},[1112,6114,591,2440],[6820,6824,6828,6832],{"plan":6821,"price":125,"features":6822,"notes":6823},"Community (Self-host)","全部功能 + 不限执行 + 不限 workflow + Sustainable Use License","VPS $5-10\u002F月",{"plan":5110,"price":6825,"features":6826,"notes":6827},"~€20\u002F月","2,500 executions + 5 workflow + 基础集成","试水 \u002F 小团队",{"plan":5121,"price":6829,"features":6830,"notes":6831},"~€50\u002F月","更高 executions + 团队协作 + 高级特性","中型团队",{"plan":728,"price":146,"features":6833,"notes":6834},"SSO + audit + LDAP + 私有部署 + SLA","大客户","Self-host 免费 \u002F Cloud Starter ~€20·月 (2.5k executions) \u002F Pro \u002F Business \u002F Enterprise 阶梯",[6837],"onboarding\u002Fn8n-ollama-automation",{"power":596,"ux":595,"price":596,"cn_support":563,"stability":596},{"title":2321,"description":6802},"n8n 评测 2026：开源自托管自动化平台，AI Agent 工作流首选",[6842,6844,6846,6848],{"name":6843,"url":6763,"accessed":2498},"n8n 官网",{"name":6845,"url":6770,"accessed":2498},"AutomationByExperts — n8n 2026 200k users 5x ARR",{"name":6847,"url":6777,"accessed":2498},"Tutorials Technology — n8n + AI on Linux 2026",{"name":6849,"url":6784,"accessed":2498},"Northflank — n8n Self-host Pricing 2026","tools\u002Fagent\u002Fplatform\u002Fn8n","自托管自动化平台：200k+ 用户 + 70 LangChain 节点 + MCP 原生 + Ollama 集成",[610,2494,6853,5528,3934,1112,2321],"automation","Zapier \u002F Make 的开源替代——20 步工作流和 2 步成本一样（自托管）。AI Agent + LangChain 节点让它在 2026 成为开发者首选自动化平台。要 8000+ 现成集成 + 极简上手用 Zapier；要中等复杂 + 不自托管用 Make。","Y5uedkjOvG7WLZTjkH2_rw8c5QipFprpPc7QrzBLivw",{"id":6857,"title":1053,"alternatives":6858,"api_compatible":15,"body":6859,"category":577,"chinese_friendly":595,"cover":7339,"description":7340,"domestic":580,"extension":581,"faq":15,"free":580,"github":7324,"languages":7341,"lastVerified":584,"meta":7342,"models":15,"navigation":586,"notSuitable":15,"opensource":586,"path":7343,"pillar":588,"platforms":7344,"priceTable":15,"pricing":5516,"published":593,"relatedPlaybooks":15,"relatedReviews":15,"score":7345,"self_host":580,"seo":7346,"seoTitle":7347,"slug":7348,"sources":7349,"stem":7352,"suitable":15,"tagline":7353,"tags":7354,"updated":584,"verdict":7356,"website":7318,"__hash__":7357},"tools\u002Ftools\u002Fagent\u002Fplatform\u002Fragflow.md",[618,619,1105],{"type":17,"value":6860,"toc":7326},[6861,6863,6866,6869,6871,6924,6926,6967,6971,6973,6979,6999,7003,7026,7028,7057,7059,7185,7187,7235,7237,7269,7271,7277,7283,7289,7295,7297,7306,7308,7312],[20,6862,23],{"id":22},[25,6864,6865],{},"RAGFlow 是 InfiniFlow（中国团队）出品的开源 RAG 引擎（Apache 2.0），核心卖点是深度文档解析 + 高召回率切片 + 引用溯源。支持 PDF \u002F Word \u002F Excel \u002F PPT \u002F 图片 \u002F 扫描件，内置 OCR + 版面分析 + 表格识别，切片质量远超通用 RAG 方案。Docker 自托管 + Cloud 云端，中文支持好（文档 \u002F UI \u002F 社区）。",[25,6867,6868],{},"适合：需要精准文档问答的企业知识库、复杂文档（表格 \u002F 图文 \u002F 扫描件）场景、中文 RAG 需求、对召回率要求高的业务。不适合：需要复杂 Agent 编排（用 Dify）、资源有限的小服务器、需要精美 UI 的 C 端产品。",[20,6870,33],{"id":33},[35,6872,6873,6879,6885,6891,6897,6903,6908,6913,6919],{},[38,6874,6875,6878],{},[41,6876,6877],{},"深度文档解析","：PDF \u002F Word \u002F Excel \u002F PPT \u002F 图片 \u002F 扫描件，版面分析 + 表格识别",[38,6880,6881,6884],{},[41,6882,6883],{},"OCR 引擎","：内置 PaddleOCR \u002F DeepDOC，支持中英文扫描件识别",[38,6886,6887,6890],{},[41,6888,6889],{},"智能切片","：基于版面分析的语义切片，保留段落 \u002F 表格 \u002F 标题结构",[38,6892,6893,6896],{},[41,6894,6895],{},"高召回率","：混合检索（全文 + 向量）+ 重排序（Rerank），召回精度高",[38,6898,6899,6902],{},[41,6900,6901],{},"引用溯源","：回答标注来源文档 + 页码 + 原文片段，可验证",[38,6904,6905,6907],{},[41,6906,5067],{},"：OpenAI \u002F Claude \u002F Ollama \u002F 通义千问 \u002F 智谱 \u002F 月之暗面",[38,6909,6910,6912],{},[41,6911,662],{},"：Elasticsearch \u002F Infinity（自研）\u002F Chroma",[38,6914,6915,6918],{},[41,6916,6917],{},"知识库管理","：多知识库 + 文档分类 + 解析状态监控",[38,6920,6921,6923],{},[41,6922,686],{},"：完整 REST API + SDK，可集成到外部系统",[20,6925,95],{"id":95},[97,6927,6928,6938],{},[100,6929,6930],{},[103,6931,6932,6934,6936],{},[106,6933,108],{},[106,6935,95],{},[106,6937,113],{},[115,6939,6940,6948,6958],{},[103,6941,6942,6944,6946],{},[120,6943,122],{},[120,6945,125],{},[120,6947,5105],{},[103,6949,6950,6952,6955],{},[120,6951,717],{},[120,6953,6954],{},"按量付费",[120,6956,6957],{},"托管服务，免运维",[103,6959,6960,6962,6964],{},[120,6961,728],{},[120,6963,146],{},[120,6965,6966],{},"私有部署 + 技术支持 + 定制",[151,6968,6969],{},[25,6970,155],{},[20,6972,159],{"id":158},[151,6974,6975],{},[25,6976,164,6977],{},[41,6978,167],{},[35,6980,6981,6984,6987,6990,6993,6996],{},[38,6982,6983],{},"文档解析质量在开源 RAG 中最强——复杂表格、多栏排版、图文混排都能正确识别",[38,6985,6986],{},"扫描件 OCR 效果好，中文印刷体识别准确率高",[38,6988,6989],{},"引用溯源到页码 + 原文片段，回答可信度高",[38,6991,6992],{},"混合检索 + Rerank 召回精度明显优于纯向量检索",[38,6994,6995],{},"中国团队出品，中文文档和社区支持好，Issue 响应快",[38,6997,6998],{},"支持通义千问 \u002F 智谱 \u002F 月之暗面等国产模型，国内场景适配好",[25,7000,7001],{},[41,7002,195],{},[35,7004,7005,7008,7011,7014,7017,7020,7023],{},[38,7006,7007],{},"资源消耗大——Elasticsearch + Redis + MinIO + RAGFlow 本身，至少 16GB 内存",[38,7009,7010],{},"部署较重，Docker Compose 起来 5+ 容器，配置复杂",[38,7012,7013],{},"大文件解析慢——100 页 PDF 解析 + 切片可能 5-10 分钟",[38,7015,7016],{},"UI 仍有粗糙处，文档管理界面交互不够流畅",[38,7018,7019],{},"Agent 能力弱——RAG 问答是强项，复杂工具调用 \u002F 多步推理不如 Dify",[38,7021,7022],{},"解析失败的重试机制不完善，偶尔卡在 parsing 状态",[38,7024,7025],{},"版本迭代快，升级需注意数据迁移",[20,7027,221],{"id":221},[223,7029,7030,7033,7039,7045,7051,7054],{},[38,7031,7032],{},"系统准备：确保 16GB+ 内存 + Docker + Docker Compose",[38,7034,7035,7036],{},"克隆仓库：",[228,7037,7038],{},"git clone https:\u002F\u002Fgithub.com\u002Finfiniflow\u002Fragflow.git",[38,7040,7041,7042],{},"启动服务：",[228,7043,7044],{},"cd ragflow\u002Fdocker && docker compose up -d",[38,7046,5209,7047,7050],{},[228,7048,7049],{},"http:\u002F\u002Flocalhost:80","，注册管理员账号",[38,7052,7053],{},"配置模型：Settings → Model Providers 添加 LLM + Embedding + Rerank",[38,7055,7056],{},"创建知识库 → 上传文档 → 等待解析完成 → 开始问答",[20,7058,266],{"id":266},[97,7060,7061,7075],{},[100,7062,7063],{},[103,7064,7065,7067,7069,7071,7073],{},[106,7066,275],{},[106,7068,1053],{},[106,7070,286],{},[106,7072,840],{},[106,7074,616],{},[115,7076,7077,7092,7105,7119,7132,7145,7157,7170],{},[103,7078,7079,7082,7085,7087,7089],{},[120,7080,7081],{},"文档解析",[120,7083,7084],{},"✅ 最强",[120,7086,855],{},[120,7088,897],{},[120,7090,7091],{},"弱",[103,7093,7094,7097,7099,7101,7103],{},[120,7095,7096],{},"表格识别",[120,7098,370],{},[120,7100,381],{},[120,7102,370],{},[120,7104,381],{},[103,7106,7107,7110,7112,7114,7117],{},[120,7108,7109],{},"OCR",[120,7111,361],{},[120,7113,381],{},[120,7115,7116],{},"需配置",[120,7118,7116],{},[103,7120,7121,7124,7126,7128,7130],{},[120,7122,7123],{},"召回精度",[120,7125,299],{},[120,7127,299],{},[120,7129,299],{},[120,7131,855],{},[103,7133,7134,7136,7139,7141,7143],{},[120,7135,6901],{},[120,7137,7138],{},"✅ 页码+片段",[120,7140,370],{},[120,7142,370],{},[120,7144,370],{},[103,7146,7147,7149,7151,7153,7155],{},[120,7148,674],{},[120,7150,7091],{},[120,7152,897],{},[120,7154,855],{},[120,7156,894],{},[103,7158,7159,7162,7164,7166,7168],{},[120,7160,7161],{},"资源消耗",[120,7163,299],{},[120,7165,855],{},[120,7167,855],{},[120,7169,296],{},[103,7171,7172,7175,7178,7180,7182],{},[120,7173,7174],{},"中文支持",[120,7176,7177],{},"✅ 优秀",[120,7179,1386],{},[120,7181,1386],{},[120,7183,7184],{},"一般",[20,7186,405],{"id":405},[35,7188,7189,7195,7205,7211,7217,7223,7229],{},[38,7190,7191,7194],{},[41,7192,7193],{},"资源一定要够","：低于 16GB 内存别部署，ES + Redis + MinIO 都吃内存",[38,7196,7197,7200,7201,7204],{},[41,7198,7199],{},"Elasticsearch 配置","：默认 JVM 堆偏小，大知识库调 ",[228,7202,7203],{},"ES_JAVA_OPTS"," 到 4-8GB",[38,7206,7207,7210],{},[41,7208,7209],{},"大文件拆分上传","：超过 100 页的 PDF 拆成小文件，解析更稳定",[38,7212,7213,7216],{},[41,7214,7215],{},"解析失败检查格式","：加密 PDF \u002F 损坏文件会卡住，上传前检查",[38,7218,7219,7222],{},[41,7220,7221],{},"Rerank 模型别省","：召回精度提升的关键，用 bge-reranker 或 Cohere Rerank",[38,7224,7225,7228],{},[41,7226,7227],{},"不要当 Agent 平台用","：RAG 问答是核心，复杂工具调用上 Dify",[38,7230,7231,7234],{},[41,7232,7233],{},"定期备份","：ES 数据 + MinIO 文件，升级前完整快照",[20,7236,457],{"id":456},[35,7238,7239,7242,7245,7248,7251,7254,7257,7260,7263,7266],{},[38,7240,7241],{},"✅ 需要精准文档问答的企业知识库",[38,7243,7244],{},"✅ 复杂文档（表格 \u002F 图文 \u002F 扫描件）RAG 场景",[38,7246,7247],{},"✅ 中文 RAG 需求（国产模型 + 中文 OCR）",[38,7249,7250],{},"✅ 对召回率和引用溯源要求高的业务",[38,7252,7253],{},"✅ 有运维能力的团队私有化部署",[38,7255,7256],{},"❌ 需要复杂 Agent 编排（用 Dify）",[38,7258,7259],{},"❌ 资源有限的小服务器（至少 16GB 内存）",[38,7261,7262],{},"❌ 需要精美 C 端 UI 的产品",[38,7264,7265],{},"❌ 无运维能力的团队（用 Cloud 版或 FastGPT）",[38,7267,7268],{},"❌ 纯英文简单文档场景（AnythingLLM 更轻量）",[20,7270,493],{"id":492},[25,7272,7273,7276],{},[41,7274,7275],{},"Q: RAGFlow 和 Dify 怎么选？","\nA: RAGFlow 专注 RAG——文档解析 + 检索精度 + 引用溯源是核心强项，适合文档密集型知识库。Dify 是完整 AI 应用平台——工作流 + Agent + RAG + API 管理，功能更全。纯文档问答选 RAGFlow，构建 AI 应用选 Dify，两者也可配合使用。",[25,7278,7279,7282],{},[41,7280,7281],{},"Q: 部署需要什么配置？","\nA: 最低 16GB 内存 + 4 核 CPU + 50GB 磁盘。生产环境建议 32GB 内存 + 8 核 + SSD。Elasticsearch 是内存大户，知识库文档量大时 ES JVM 堆需 8GB+。如果资源有限，考虑用 Infinity（RAGFlow 自研向量库）替代 ES。",[25,7284,7285,7288],{},[41,7286,7287],{},"Q: 支持中文 OCR 吗？","\nA: 支持。内置 PaddleOCR + DeepDOC 引擎，中文印刷体识别准确率高。手写体效果一般，复杂背景的扫描件建议预处理（去噪 \u002F 矫正）后再上传。OCR 默认开启，可在解析模板中配置。",[25,7290,7291,7294],{},[41,7292,7293],{},"Q: 和 FastGPT 比 RAG 精度如何？","\nA: 两者 RAG 精度都属第一梯队。RAGFlow 的优势在文档解析——复杂表格、多栏版面、图文混排的识别更准确，切片质量更高。FastGPT 的优势在工作流编排和知识库管理 UI 更成熟。文档解析要求高选 RAGFlow，流程管理要求高选 FastGPT。",[20,7296,520],{"id":520},[25,7298,7299,527,7302,527,7304],{},[524,7300,616],{"href":7301},"\u002Fagent\u002Fplatform\u002Fanythingllm.html",[524,7303,531],{"href":530},[524,7305,535],{"href":534},[20,7307,538],{"id":538},[151,7309,7310],{},[25,7311,543],{},[35,7313,7314,7320],{},[38,7315,7316],{},[524,7317,553],{"href":7318,"rel":7319},"https:\u002F\u002Fragflow.io",[552],[38,7321,7322],{},[524,7323,560],{"href":7324,"rel":7325},"https:\u002F\u002Fgithub.com\u002Finfiniflow\u002Fragflow",[552],{"title":562,"searchDepth":563,"depth":563,"links":7327},[7328,7329,7330,7331,7332,7333,7334,7335,7336,7337,7338],{"id":22,"depth":566,"text":23},{"id":33,"depth":566,"text":33},{"id":95,"depth":566,"text":95},{"id":158,"depth":566,"text":159},{"id":221,"depth":566,"text":221},{"id":266,"depth":566,"text":266},{"id":405,"depth":566,"text":405},{"id":456,"depth":566,"text":457},{"id":492,"depth":566,"text":493},{"id":520,"depth":566,"text":520},{"id":538,"depth":566,"text":538},"\u002Fimg\u002Ftools\u002Fragflow.webp","RAGFlow 真实评测：InfiniFlow 出品的开源 RAG 引擎（Apache 2.0 协议），深度文档解析（PDF\u002FWord\u002FExcel\u002F图片）+ 高召回率切片 + 引用溯源。支持 Docker 自托管，适合需要精准文档问答和知识库检索的企业场景。",[583,2426],{},"\u002Ftools\u002Fagent\u002Fplatform\u002Fragflow",[590,591],{"power":595,"ux":563,"price":596,"cn_support":595,"stability":563},{"title":1053,"description":7340},"RAGFlow - 开源 RAG 引擎评测与部署 | AIHO","agent\u002Fplatform\u002Fragflow",[7350,7351],{"title":553,"url":7318},{"title":560,"url":7324},"tools\u002Fagent\u002Fplatform\u002Fragflow","开源 RAG 引擎，深度文档解析 + 高召回率",[606,610,1113,7355,1112],"document-parsing","需要精准文档解析和高召回率 RAG 的企业场景首选，深度文档解析（复杂表格\u002F版面\u002FOCR）+ 引用溯源能力在开源 RAG 引擎中最强，中国团队出品中文支持好，但部署资源要求高、Agent 能力弱、UI 仍需打磨。","bLpjRG4MMrBFYyeFSz4XzLMwVql0SyGD3sB3Tr5g38g",{"id":7359,"title":7360,"alternatives":7361,"api_compatible":15,"body":7362,"category":577,"chinese_friendly":596,"cover":7844,"description":7845,"domestic":580,"extension":581,"faq":7846,"free":580,"github":15,"languages":7859,"lastVerified":15,"meta":7860,"models":15,"navigation":586,"notSuitable":15,"opensource":580,"path":7861,"pillar":588,"platforms":7862,"priceTable":7866,"pricing":7878,"published":6129,"relatedPlaybooks":7879,"relatedReviews":15,"score":7881,"self_host":580,"seo":7882,"seoTitle":7883,"slug":1593,"sources":7884,"stem":7893,"suitable":15,"tagline":7894,"tags":7895,"updated":2498,"verdict":7901,"website":7809,"__hash__":7902},"tools\u002Ftools\u002Fagent\u002Fplatform\u002Fyuanqi.md","腾讯元器",[12,14,5534],{"type":17,"value":7363,"toc":7832},[7364,7366,7369,7372,7374,7448,7450,7469,7474,7478,7482,7508,7512,7538,7542,7565,7567,7699,7701,7756,7758,7784,7786,7800,7802],[20,7365,23],{"id":22},[25,7367,7368],{},"腾讯元器（yuanqi.tencent.com）是腾讯官方零代码智能体平台，2024 推出、2026-01 完成系统升级。差异点：『人人可用、零代码、零基础』+ 公众号扫码一键授权（历史群发文章 → 智能体知识库）+ 深度整合 LLM + RAG + 工作流 + 多 Agent 协作 + 一键发布到微信客服 \u002F 公众号 \u002F 服务号 \u002F 企业微信 \u002F 应用宝 + 打通微信支付 MCP 闭环 + 智能体广场 \u002F 模板 \u002F 插件广场 + 混元 + DeepSeek 双底座。",[25,7370,7371],{},"适合：微信公众号博主 \u002F 国内自媒体 + 想做粉丝 7×24 问答；政企客服 \u002F 政务 \u002F 民生场景；电商品牌 + 微信生态 + 微信支付闭环；非技术运营人员 + 零代码上手。不适合：海外用户 \u002F 跨平台分发（用 Coze 海外版 \u002F Dify）；需要自托管 \u002F 数据零云（用 Dify）；纯通用 reasoning \u002F 学术 agent；要深度 API + 工程化（用 Langflow \u002F LangChain）。",[20,7373,33],{"id":33},[35,7375,7376,7382,7388,7394,7400,7406,7412,7418,7424,7430,7436,7442],{},[38,7377,7378,7381],{},[41,7379,7380],{},"公众号一键授权","：扫码绑定 → 历史文章变知识库",[38,7383,7384,7387],{},[41,7385,7386],{},"零代码创建","：自然语言描述 → 智能体 + 工作流",[38,7389,7390,7393],{},[41,7391,7392],{},"混元 + DeepSeek 双底座","：选模型 + 备案合规",[38,7395,7396,7399],{},[41,7397,7398],{},"知识库 RAG","：文章 \u002F PDF \u002F 网页 \u002F 自定义文本",[38,7401,7402,7405],{},[41,7403,7404],{},"工作流引擎","：可视化编排多步执行",[38,7407,7408,7411],{},[41,7409,7410],{},"多 Agent 协作","：子 agent 调子 agent 完成复杂任务",[38,7413,7414,7417],{},[41,7415,7416],{},"插件广场","：搜索 \u002F 天气 \u002F 翻译 \u002F 图片生成等",[38,7419,7420,7423],{},[41,7421,7422],{},"智能体广场 \u002F 模板","：clone 业内成熟智能体",[38,7425,7426,7429],{},[41,7427,7428],{},"全微信生态发布","：客服 \u002F 公众号 \u002F 服务号 \u002F 企业微信 \u002F 应用宝",[38,7431,7432,7435],{},[41,7433,7434],{},"微信支付 MCP","：智能体可直接发起支付闭环",[38,7437,7438,7441],{},[41,7439,7440],{},"企业版能力","：私有部署 + SSO + 定制",[38,7443,7444,7447],{},[41,7445,7446],{},"小助手 \u002F 客服助手 \u002F IP 分身 \u002F 角色陪聊","等内置场景",[20,7449,95],{"id":95},[35,7451,7452,7458,7463],{},[38,7453,7454,7457],{},[41,7455,7456],{},"免费创建","：¥0；创建智能体 + 公众号绑定 + 知识库 + 工作流",[38,7459,7460,7462],{},[41,7461,6954],{},"：混元 \u002F DeepSeek token + 知识库检索 + 工作流执行 + 插件调用按次数",[38,7464,7465,7468],{},[41,7466,7467],{},"企业级方案","：联系商务；私有部署 + SSO + 定制 + SLA",[151,7470,7471],{},[25,7472,7473],{},"实际成本：个人公众号博主一月几十块 token 钱足够跑数千次 7×24 问答；企业客服需求要走商务报价。",[20,7475,7477],{"id":7476},"实测公众号自媒体-政企客服","实测（公众号自媒体 + 政企客服）",[25,7479,7480],{},[41,7481,167],{},[35,7483,7484,7487,7490,7493,7496,7499,7502,7505],{},[38,7485,7486],{},"公众号扫码授权 + 历史文章入库自动化程度极高，5 分钟跑通",[38,7488,7489],{},"混元 + DeepSeek 中文效果好 + 国内备案合规",[38,7491,7492],{},"微信生态全链路：粉丝问 → 智能体答 → 推荐产品 → 微信支付 闭环",[38,7494,7495],{},"智能体广场可看到童爸育儿（2000+ 篇文章库）\u002F 飞哥律界 等真实案例",[38,7497,7498],{},"政务 \u002F 央企（钦州政务 \u002F 中国石化『小石头』）有正式落地",[38,7500,7501],{},"多 Agent 协作架构让复杂业务场景可拆分",[38,7503,7504],{},"零代码运营人员可独立完成，不依赖技术",[38,7506,7507],{},"智能体模板让冷启动快 10x",[25,7509,7510],{},[41,7511,195],{},[35,7513,7514,7517,7520,7523,7526,7529,7532,7535],{},[38,7515,7516],{},"新旧版双底层，旧版用户迁移有摩擦",[38,7518,7519],{},"海外用户访问受限，跨海外业务用不上",[38,7521,7522],{},"工作流编排能力不如 Dify \u002F Coze 强（少高级控制流）",[38,7524,7525],{},"多 Agent 编排 UI 学习曲线",[38,7527,7528],{},"知识库切片粒度自定义有限",[38,7530,7531],{},"国产模型 reasoning 能力比 GPT-5 \u002F Claude 仍有差距",[38,7533,7534],{},"自托管 \u002F 数据私有化不可行（除非企业版私有部署）",[38,7536,7537],{},"海外大模型接入限制（混元 \u002F DeepSeek 为主）",[20,7539,7541],{"id":7540},"上手公众号博主-5-分钟","上手（公众号博主 5 分钟）",[223,7543,7544,7547,7550,7553,7556,7559,7562],{},[38,7545,7546],{},"yuanqi.tencent.com → 微信扫码登录",[38,7548,7549],{},"新建智能体 → 名字 + 头像 + 描述",[38,7551,7552],{},"知识库 → 绑定公众号 → 扫码授权 → 等历史文章入库（几分钟）",[38,7554,7555],{},"角色设定 → 例：『你是 X 公众号官方助手，基于知识库回答问题』",[38,7557,7558],{},"测试对话 → 验证答案命中文章",[38,7560,7561],{},"发布 → 选择渠道（公众号菜单 \u002F 客服）",[38,7563,7564],{},"粉丝在公众号点击菜单 → 进入 AI 对话",[20,7566,266],{"id":266},[97,7568,7569,7584],{},[100,7570,7571],{},[103,7572,7573,7575,7577,7579,7581],{},[106,7574,275],{},[106,7576,2172],{},[106,7578,1591],{},[106,7580,286],{},[106,7582,7583],{},"文心智能体",[115,7585,7586,7598,7612,7628,7642,7655,7672,7685],{},[103,7587,7588,7590,7592,7594,7596],{},[120,7589,7380],{},[120,7591,5873],{},[120,7593,381],{},[120,7595,381],{},[120,7597,5890],{},[103,7599,7600,7603,7606,7608,7610],{},[120,7601,7602],{},"微信支付闭环",[120,7604,7605],{},"✅ MCP",[120,7607,381],{},[120,7609,381],{},[120,7611,381],{},[103,7613,7614,7617,7620,7623,7625],{},[120,7615,7616],{},"多渠道发布",[120,7618,7619],{},"微信生态",[120,7621,7622],{},"多平台 + 海外",[120,7624,3291],{},[120,7626,7627],{},"百度生态",[103,7629,7630,7633,7635,7638,7640],{},[120,7631,7632],{},"海外可用",[120,7634,381],{},[120,7636,7637],{},"✅ Coze.com",[120,7639,370],{},[120,7641,381],{},[103,7643,7644,7646,7649,7651,7653],{},[120,7645,3291],{},[120,7647,7648],{},"❌（企业版才有）",[120,7650,381],{},[120,7652,4571],{},[120,7654,381],{},[103,7656,7657,7660,7663,7666,7669],{},[120,7658,7659],{},"LLM 底座",[120,7661,7662],{},"混元 + DeepSeek",[120,7664,7665],{},"多家",[120,7667,7668],{},"多家 + 自托管",[120,7670,7671],{},"文心一言",[103,7673,7674,7676,7679,7681,7683],{},[120,7675,4617],{},[120,7677,7678],{},"✅ 基础",[120,7680,5873],{},[120,7682,5873],{},[120,7684,370],{},[103,7686,7687,7689,7692,7694,7697],{},[120,7688,5347],{},[120,7690,7691],{},"公众号 + 国内",[120,7693,7622],{},[120,7695,7696],{},"自托管 + 企业",[120,7698,7627],{},[20,7700,405],{"id":405},[35,7702,7703,7709,7715,7721,7727,7732,7738,7744,7750],{},[38,7704,7705,7708],{},[41,7706,7707],{},"新版优先","：旧版功能停止演进，新建议直接新版",[38,7710,7711,7714],{},[41,7712,7713],{},"公众号文章质量","：知识库 = 文章质量，垃圾文章入库 = 垃圾回答",[38,7716,7717,7720],{},[41,7718,7719],{},"测试对话","：发布前测 20 个常见问题，验证答案准确",[38,7722,7723,7726],{},[41,7724,7725],{},"兜底回复","：知识库命中失败要有兜底引导（『请联系人工』）",[38,7728,7729,7731],{},[41,7730,7602],{},"：用户体验要清晰，支付链路要可中断",[38,7733,7734,7737],{},[41,7735,7736],{},"token 预算","：开通调用前看模型计价 + 设上限",[38,7739,7740,7743],{},[41,7741,7742],{},"海外用户","：访问可能不稳，业务覆盖海外要走 Coze",[38,7745,7746,7749],{},[41,7747,7748],{},"政务 \u002F 央企合规","：用企业版私有部署 + 备案",[38,7751,7752,7755],{},[41,7753,7754],{},"多 Agent 不要过度","：复杂业务先单 agent 跑通，再拆分",[20,7757,457],{"id":456},[35,7759,7760,7763,7766,7769,7772,7775,7778,7781],{},[38,7761,7762],{},"✅ 微信公众号博主 \u002F 国内自媒体",[38,7764,7765],{},"✅ 政企客服 \u002F 政务 \u002F 民生场景",[38,7767,7768],{},"✅ 电商品牌 + 微信支付闭环",[38,7770,7771],{},"✅ 零代码运营 + 非技术团队",[38,7773,7774],{},"❌ 海外用户 \u002F 跨平台分发",[38,7776,7777],{},"❌ 自托管 \u002F 数据零云",[38,7779,7780],{},"❌ 纯通用 reasoning \u002F 学术 agent",[38,7782,7783],{},"❌ 深度工程化 + API-first",[20,7785,520],{"id":520},[35,7787,7788,7792,7796],{},[38,7789,7790],{},[524,7791,6745],{"href":6112},[38,7793,7794],{},[524,7795,6035],{"href":6034},[38,7797,7798],{},[524,7799,6041],{"href":6040},[20,7801,538],{"id":538},[223,7803,7804,7811,7818,7825],{},[38,7805,7806,7807],{},"腾讯元器官网 ",[524,7808,7809],{"href":7809,"rel":7810},"https:\u002F\u002Fyuanqi.tencent.com",[552],[38,7812,7813,7814],{},"腾讯元器帮助中心 - 平台介绍 + 升级说明 ",[524,7815,7816],{"href":7816,"rel":7817},"https:\u002F\u002Fyuanqi.tencent.com\u002Fguide\u002Fyuanqi-introduction",[552],[38,7819,7820,7821],{},"智能体广场（真实落地案例：童爸育儿 \u002F 飞哥律界 \u002F 钦州政务 \u002F 中国石化小石头）",[524,7822,7823],{"href":7823,"rel":7824},"https:\u002F\u002Fyuanqi.tencent.com\u002Fagent-center\u002Findex",[552],[38,7826,7827,7828],{},"AGI 空间站 — 腾讯混元 + 元器生态 ",[524,7829,7830],{"href":7830,"rel":7831},"https:\u002F\u002Fdocs.feishu.cn\u002Fv\u002Fwiki\u002FJc1TwOaAni879pksMgmcqgUNnof\u002Fa3",[552],{"title":562,"searchDepth":563,"depth":563,"links":7833},[7834,7835,7836,7837,7838,7839,7840,7841,7842,7843],{"id":22,"depth":566,"text":23},{"id":33,"depth":566,"text":33},{"id":95,"depth":566,"text":95},{"id":7476,"depth":566,"text":7477},{"id":7540,"depth":566,"text":7541},{"id":266,"depth":566,"text":266},{"id":405,"depth":566,"text":405},{"id":456,"depth":566,"text":457},{"id":520,"depth":566,"text":520},{"id":538,"depth":566,"text":538},"\u002Fimg\u002Ftools\u002Fyuanqi.webp","腾讯元器（yuanqi.tencent.com）真实评测：腾讯官方零代码智能体平台，2024 推出、2026-01 完成系统升级（新旧双版本并存）。差异点：『人人可用、零代码、零基础』面向普通用户 + 公众号扫码一键授权（历史群发文章 → 智能体知识库）+ 深度整合 LLM + RAG + 工作流引擎 + 多 Agent 协作架构 + 一键发布到微信客服 \u002F 公众号 \u002F 服务号 \u002F 企业微信 \u002F 应用宝 + 打通微信支付 MCP 闭环 + 智能体广场 \u002F 模板 \u002F 插件广场 + 混元 + DeepSeek 双底座。",[7847,7850,7853,7856],{"q":7848,"a":7849},"和扣子 Coze \u002F Dify \u002F 文心智能体怎么选？","元器的核心独门：『微信公众号一键授权 + 文章变知识库 + 发布到微信全生态 + 微信支付 MCP』。Coze（字节）强『多 LLM + Bot 商店 + 海外版』。Dify 强『LLMOps 全平台 + 可自托管』。文心智能体强『百度生态 + 搜索 + 文心一言』。微信公众号博主 \u002F 政企客户 → 元器；多平台分发 \u002F 跨海外 → Coze；自托管 \u002F 企业 → Dify；百度生态 → 文心。",{"q":7851,"a":7852},"新旧版关系？","2025-12 元器完成系统升级，新版底层逻辑重写（多 Agent 架构 + 微信支付 MCP 等新能力）。旧版创建的智能体 \u002F 知识库 \u002F 工作流暂时无法直接同步到新版（两套底层），需走迁移路径。新建议直接用新版。官方文档：『从旧版元器迁移智能体到新版元器』。",{"q":7854,"a":7855},"公众号绑定后如何运作？","公众号管理员扫码授权 → 元器拉取历史群发文章 → 自动切片 + embedding 入知识库 → 公众号菜单栏出现『智能问答』入口 → 粉丝点击进入智能体对话 → 命中知识库的问题用文章原文 + LLM 综合回答 → 命中失败走兜底回复。这套路径让自媒体几乎零成本拥有 7×24 客服 \u002F 问答机器人。",{"q":7857,"a":7858},"数据隐私和合规？","元器是腾讯官方服务 + 数据托管在腾讯云 + 合规接入备案算法（混元 \u002F DeepSeek）。对国内合规 \u002F 政务 \u002F 央企场景非常友好。但对隐私敏感 \u002F 数据零云需求 → 走 Dify 自托管 + 国产开源模型路径，不要用元器。",[2426,583],{},"\u002Ftools\u002Fagent\u002Fplatform\u002Fyuanqi",[2440,7863,7864,7865],"wechat-mp","wechat-work","wechat-pay",[7867,7870,7874],{"plan":7456,"price":2444,"features":7868,"notes":7869},"创建智能体 + 公众号绑定 + 知识库 + 工作流 + 智能体广场","模型 token 按量",{"plan":6954,"price":7871,"features":7872,"notes":7873},"按调用","混元 \u002F DeepSeek \u002F 知识库检索 \u002F 工作流执行 \u002F 插件调用","token + API 次数",{"plan":7467,"price":7875,"features":7876,"notes":7877},"联系商务","私有部署 + SSO + audit + 定制开发 + 多 agent 协作 + SLA","政企客户","免费创建 \u002F 模型调用按 token \u002F 部分能力按调用次数 \u002F 企业版联系商务",[7880],"onboarding\u002Fwechat-mp-ai-agent",{"power":595,"ux":596,"price":596,"cn_support":596,"stability":595},{"title":7360,"description":7845},"腾讯元器 AI Agent 平台评测 2026：腾讯混元大模型应用开发",[7885,7887,7889,7891],{"name":7886,"url":7809,"accessed":2498},"腾讯元器官网",{"name":7888,"url":7816,"accessed":2498},"腾讯元器帮助中心 - 平台介绍",{"name":7890,"url":7823,"accessed":2498},"智能体广场",{"name":7892,"url":7830,"accessed":2498},"AGI 空间站 — 腾讯混元 + 元器生态","tools\u002Fagent\u002Fplatform\u002Fyuanqi","腾讯零代码智能体平台——公众号一键变 AI 分身 + 混元 \u002F DeepSeek 双底座 + 微信生态全链路",[7896,7897,2493,7898,7899,7900],"agent-builder","wechat","hunyuan","deepseek","yuanqi","微信公众号博主 \u002F 国内自媒体 \u002F 政企客服首选——把公众号变 7×24 AI 分身的最低成本路径。要纯通用 agent \u002F 海外用户 \u002F 复杂 reasoning 用 Dify \u002F Coze。","ecU4jqtN9rSFz0nKLjXaXWfKQf4tCqmUcnzZ5BCUJMI",[],[],[7906,8270,8718],{"id":1118,"title":280,"alternatives":7907,"api_compatible":15,"body":7908,"category":577,"chinese_friendly":566,"cover":1570,"description":1571,"domestic":580,"extension":581,"faq":15,"free":580,"github":1555,"languages":8261,"lastVerified":584,"meta":8262,"models":15,"navigation":586,"notSuitable":15,"opensource":586,"path":1574,"pillar":588,"platforms":8263,"priceTable":15,"pricing":1576,"published":593,"relatedPlaybooks":15,"relatedReviews":15,"score":8264,"self_host":580,"seo":8265,"seoTitle":1579,"slug":13,"sources":8266,"stem":1583,"suitable":15,"tagline":1584,"tags":8269,"updated":584,"verdict":1587,"website":1549,"__hash__":1588},[599,12,618],{"type":17,"value":7909,"toc":8248},[7910,7912,7916,7918,7920,7958,7960,7962,7964,7970,7986,7990,8006,8008,8032,8034,8136,8138,8170,8172,8194,8196,8200,8210,8214,8220,8222,8230,8232,8236],[20,7911,23],{"id":22},[25,7913,1126,7914,1130],{},[228,7915,1129],{},[25,7917,1133],{},[20,7919,33],{"id":33},[35,7921,7922,7926,7930,7934,7938,7942,7946,7950,7954],{},[38,7923,7924,1143],{},[41,7925,1142],{},[38,7927,7928,1148],{},[41,7929,343],{},[38,7931,7932,1154],{},[41,7933,1153],{},[38,7935,7936,1160],{},[41,7937,1159],{},[38,7939,7940,1166],{},[41,7941,1165],{},[38,7943,7944,1172],{},[41,7945,1171],{},[38,7947,7948,1177],{},[41,7949,79],{},[38,7951,7952,1183],{},[41,7953,1182],{},[38,7955,7956,1189],{},[41,7957,1188],{},[20,7959,95],{"id":95},[25,7961,1194],{},[20,7963,159],{"id":158},[151,7965,7966],{},[25,7967,164,7968],{},[41,7969,167],{},[35,7971,7972,7974,7976,7978,7982,7984],{},[38,7973,1207],{},[38,7975,1210],{},[38,7977,1213],{},[38,7979,1216,7980,1220],{},[228,7981,1219],{},[38,7983,1223],{},[38,7985,1226],{},[25,7987,7988],{},[41,7989,195],{},[35,7991,7992,7994,7996,7998,8000,8002,8004],{},[38,7993,1235],{},[38,7995,1238],{},[38,7997,1241],{},[38,7999,1244],{},[38,8001,1247],{},[38,8003,1250],{},[38,8005,1253],{},[20,8007,221],{"id":221},[223,8009,8010,8014,8018,8022,8026,8030],{},[38,8011,8012,1263],{},[228,8013,1262],{},[38,8015,234,8016,1268],{},[228,8017,237],{},[38,8019,241,8020],{},[228,8021,1273],{},[38,8023,1276,8024],{},[228,8025,1279],{},[38,8027,1282,8028],{},[228,8029,1285],{},[38,8031,1288],{},[20,8033,266],{"id":266},[97,8035,8036,8050],{},[100,8037,8038],{},[103,8039,8040,8042,8044,8046,8048],{},[106,8041,275],{},[106,8043,280],{},[106,8045,10],{},[106,8047,283],{},[106,8049,286],{},[115,8051,8052,8064,8076,8088,8100,8112,8124],{},[103,8053,8054,8056,8058,8060,8062],{},[120,8055,1313],{},[120,8057,1316],{},[120,8059,1316],{},[120,8061,1316],{},[120,8063,1323],{},[103,8065,8066,8068,8070,8072,8074],{},[120,8067,293],{},[120,8069,299],{},[120,8071,855],{},[120,8073,299],{},[120,8075,296],{},[103,8077,8078,8080,8082,8084,8086],{},[120,8079,326],{},[120,8081,332],{},[120,8083,329],{},[120,8085,335],{},[120,8087,338],{},[103,8089,8090,8092,8094,8096,8098],{},[120,8091,343],{},[120,8093,349],{},[120,8095,346],{},[120,8097,346],{},[120,8099,354],{},[103,8101,8102,8104,8106,8108,8110],{},[120,8103,375],{},[120,8105,381],{},[120,8107,1368],{},[120,8109,381],{},[120,8111,370],{},[103,8113,8114,8116,8118,8120,8122],{},[120,8115,1377],{},[120,8117,1380],{},[120,8119,1383],{},[120,8121,1386],{},[120,8123,1386],{},[103,8125,8126,8128,8130,8132,8134],{},[120,8127,390],{},[120,8129,1395],{},[120,8131,393],{},[120,8133,1400],{},[120,8135,402],{},[20,8137,405],{"id":405},[35,8139,8140,8146,8150,8154,8158,8162,8166],{},[38,8141,8142,1412,8144,1416],{},[41,8143,1411],{},[228,8145,1415],{},[38,8147,8148,1422],{},[41,8149,1421],{},[38,8151,8152,1428],{},[41,8153,1427],{},[38,8155,8156,1434],{},[41,8157,1433],{},[38,8159,8160,1440],{},[41,8161,1439],{},[38,8163,8164,1446],{},[41,8165,1445],{},[38,8167,8168,1452],{},[41,8169,1451],{},[20,8171,457],{"id":456},[35,8173,8174,8176,8178,8180,8182,8184,8186,8188,8190,8192],{},[38,8175,1459],{},[38,8177,1462],{},[38,8179,1465],{},[38,8181,1468],{},[38,8183,1471],{},[38,8185,1474],{},[38,8187,477],{},[38,8189,1479],{},[38,8191,1482],{},[38,8193,1485],{},[20,8195,493],{"id":492},[25,8197,8198,1493],{},[41,8199,1492],{},[25,8201,8202,1499,8204,1503,8206,1507,8208,1511],{},[41,8203,1498],{},[228,8205,1502],{},[228,8207,1506],{},[228,8209,1510],{},[25,8211,8212,1517],{},[41,8213,1516],{},[25,8215,8216,1522,8218,1525],{},[41,8217,510],{},[228,8219,1510],{},[20,8221,520],{"id":520},[25,8223,8224,527,8226,527,8228],{},[524,8225,10],{"href":1532},[524,8227,531],{"href":530},[524,8229,535],{"href":534},[20,8231,538],{"id":538},[151,8233,8234],{},[25,8235,543],{},[35,8237,8238,8243],{},[38,8239,8240],{},[524,8241,553],{"href":1549,"rel":8242},[552],[38,8244,8245],{},[524,8246,560],{"href":1555,"rel":8247},[552],{"title":562,"searchDepth":563,"depth":563,"links":8249},[8250,8251,8252,8253,8254,8255,8256,8257,8258,8259,8260],{"id":22,"depth":566,"text":23},{"id":33,"depth":566,"text":33},{"id":95,"depth":566,"text":95},{"id":158,"depth":566,"text":159},{"id":221,"depth":566,"text":221},{"id":266,"depth":566,"text":266},{"id":405,"depth":566,"text":405},{"id":456,"depth":566,"text":457},{"id":492,"depth":566,"text":493},{"id":520,"depth":566,"text":520},{"id":538,"depth":566,"text":538},[583],{},[590,591],{"power":595,"ux":566,"price":596,"cn_support":566,"stability":563},{"title":280,"description":1571},[8267,8268],{"title":553,"url":1549},{"title":560,"url":1555},[606,607,608,1586,609,610],{"id":5532,"title":843,"alternatives":8271,"api_compatible":15,"body":8272,"category":577,"chinese_friendly":563,"cover":6094,"description":6095,"domestic":580,"extension":581,"faq":8696,"free":580,"github":15,"languages":8701,"lastVerified":15,"meta":8702,"models":15,"navigation":586,"notSuitable":15,"opensource":586,"path":6112,"pillar":588,"platforms":8703,"priceTable":8704,"pricing":6128,"published":6129,"relatedPlaybooks":8709,"relatedReviews":15,"score":8710,"self_host":586,"seo":8711,"seoTitle":6134,"slug":12,"sources":8712,"stem":6145,"suitable":15,"tagline":6146,"tags":8717,"updated":2498,"verdict":6150,"website":6138,"__hash__":6151},[14,5534,5535],{"type":17,"value":8273,"toc":8684},[8274,8276,8278,8280,8282,8332,8334,8352,8356,8358,8362,8380,8384,8402,8404,8448,8450,8466,8468,8582,8584,8622,8624,8642,8644,8658,8660,8682],[20,8275,23],{"id":22},[25,8277,5542],{},[25,8279,5545],{},[20,8281,33],{"id":33},[35,8283,8284,8288,8292,8296,8300,8304,8308,8312,8316,8320,8324,8328],{},[38,8285,8286,5555],{},[41,8287,5554],{},[38,8289,8290,5561],{},[41,8291,5560],{},[38,8293,8294,5567],{},[41,8295,5566],{},[38,8297,8298,5573],{},[41,8299,5572],{},[38,8301,8302,5579],{},[41,8303,5578],{},[38,8305,8306,5585],{},[41,8307,5584],{},[38,8309,8310,5591],{},[41,8311,5590],{},[38,8313,8314,5597],{},[41,8315,5596],{},[38,8317,8318,5603],{},[41,8319,5602],{},[38,8321,8322,5609],{},[41,8323,5608],{},[38,8325,8326,5615],{},[41,8327,5614],{},[38,8329,8330,5621],{},[41,8331,5620],{},[20,8333,95],{"id":95},[35,8335,8336,8340,8344,8348],{},[38,8337,8338,5631],{},[41,8339,5630],{},[38,8341,8342,5637],{},[41,8343,5636],{},[38,8345,8346,5643],{},[41,8347,5642],{},[38,8349,8350,5648],{},[41,8351,728],{},[151,8353,8354],{},[25,8355,5653],{},[20,8357,5657],{"id":5656},[25,8359,8360],{},[41,8361,167],{},[35,8363,8364,8366,8368,8370,8372,8374,8376,8378],{},[38,8365,5666],{},[38,8367,5669],{},[38,8369,5672],{},[38,8371,5675],{},[38,8373,5678],{},[38,8375,5681],{},[38,8377,5684],{},[38,8379,5687],{},[25,8381,8382],{},[41,8383,195],{},[35,8385,8386,8388,8390,8392,8394,8396,8398,8400],{},[38,8387,5696],{},[38,8389,5699],{},[38,8391,5702],{},[38,8393,5705],{},[38,8395,5708],{},[38,8397,5711],{},[38,8399,5714],{},[38,8401,5717],{},[20,8403,221],{"id":221},[3185,8405,8406],{"className":3187,"code":5722,"language":3189,"meta":562,"style":562},[228,8407,8408,8412,8420,8428,8432,8436],{"__ignoreMap":562},[3193,8409,8410],{"class":3195,"line":3196},[3193,8411,5729],{"class":3245},[3193,8413,8414,8416,8418],{"class":3195,"line":566},[3193,8415,5734],{"class":3199},[3193,8417,5737],{"class":3203},[3193,8419,5740],{"class":3203},[3193,8421,8422,8424,8426],{"class":3195,"line":563},[3193,8423,5745],{"class":3199},[3193,8425,5748],{"class":3203},[3193,8427,5751],{"class":3245},[3193,8429,8430],{"class":3195,"line":595},[3193,8431,3391],{"emptyLinePlaceholder":586},[3193,8433,8434],{"class":3195,"line":596},[3193,8435,5760],{"class":3245},[3193,8437,8438,8440,8442,8444,8446],{"class":3195,"line":3388},[3193,8439,591],{"class":3199},[3193,8441,5748],{"class":3203},[3193,8443,5769],{"class":3212},[3193,8445,5772],{"class":3203},[3193,8447,5775],{"class":3203},[25,8449,5778],{},[223,8451,8452,8454,8456,8458,8460,8462,8464],{},[38,8453,5783],{},[38,8455,5786],{},[38,8457,5789],{},[38,8459,5792],{},[38,8461,5795],{},[38,8463,5798],{},[38,8465,5801],{},[20,8467,266],{"id":266},[97,8469,8470,8484],{},[100,8471,8472],{},[103,8473,8474,8476,8478,8480,8482],{},[106,8475,275],{},[106,8477,843],{},[106,8479,286],{},[106,8481,2321],{},[106,8483,531],{},[115,8485,8486,8498,8510,8522,8534,8546,8558,8570],{},[103,8487,8488,8490,8492,8494,8496],{},[120,8489,5826],{},[120,8491,5829],{},[120,8493,5832],{},[120,8495,5835],{},[120,8497,5838],{},[103,8499,8500,8502,8504,8506,8508],{},[120,8501,2018],{},[120,8503,4580],{},[120,8505,5847],{},[120,8507,5850],{},[120,8509,4580],{},[103,8511,8512,8514,8516,8518,8520],{},[120,8513,3291],{},[120,8515,5859],{},[120,8517,5862],{},[120,8519,5862],{},[120,8521,370],{},[103,8523,8524,8526,8528,8530,8532],{},[120,8525,321],{},[120,8527,5873],{},[120,8529,370],{},[120,8531,370],{},[120,8533,370],{},[103,8535,8536,8538,8540,8542,8544],{},[120,8537,5884],{},[120,8539,5887],{},[120,8541,5890],{},[120,8543,5893],{},[120,8545,5893],{},[103,8547,8548,8550,8552,8554,8556],{},[120,8549,5900],{},[120,8551,370],{},[120,8553,370],{},[120,8555,5890],{},[120,8557,370],{},[103,8559,8560,8562,8564,8566,8568],{},[120,8561,5913],{},[120,8563,5916],{},[120,8565,5919],{},[120,8567,5922],{},[120,8569,5925],{},[103,8571,8572,8574,8576,8578,8580],{},[120,8573,5347],{},[120,8575,5932],{},[120,8577,5935],{},[120,8579,5277],{},[120,8581,5940],{},[20,8583,405],{"id":405},[35,8585,8586,8590,8594,8598,8602,8606,8610,8614,8618],{},[38,8587,8588,5950],{},[41,8589,5949],{},[38,8591,8592,5956],{},[41,8593,5955],{},[38,8595,8596,5962],{},[41,8597,5961],{},[38,8599,8600,5968],{},[41,8601,5967],{},[38,8603,8604,5974],{},[41,8605,5973],{},[38,8607,8608,5980],{},[41,8609,5979],{},[38,8611,8612,5986],{},[41,8613,5985],{},[38,8615,8616,5991],{},[41,8617,3577],{},[38,8619,8620,5997],{},[41,8621,5996],{},[20,8623,457],{"id":456},[35,8625,8626,8628,8630,8632,8634,8636,8638,8640],{},[38,8627,6004],{},[38,8629,6007],{},[38,8631,6010],{},[38,8633,6013],{},[38,8635,6016],{},[38,8637,6019],{},[38,8639,6022],{},[38,8641,6025],{},[20,8643,520],{"id":520},[35,8645,8646,8650,8654],{},[38,8647,8648],{},[524,8649,6035],{"href":6034},[38,8651,8652],{},[524,8653,6041],{"href":6040},[38,8655,8656],{},[524,8657,6047],{"href":6046},[20,8659,538],{"id":538},[223,8661,8662,8667,8672,8677],{},[38,8663,6054,8664],{},[524,8665,6057],{"href":6057,"rel":8666},[552],[38,8668,6061,8669],{},[524,8670,6064],{"href":6064,"rel":8671},[552],[38,8673,6068,8674],{},[524,8675,6071],{"href":6071,"rel":8676},[552],[38,8678,6075,8679],{},[524,8680,6078],{"href":6078,"rel":8681},[552],[3864,8683,4924],{},{"title":562,"searchDepth":563,"depth":563,"links":8685},[8686,8687,8688,8689,8690,8691,8692,8693,8694,8695],{"id":22,"depth":566,"text":23},{"id":33,"depth":566,"text":33},{"id":95,"depth":566,"text":95},{"id":5656,"depth":566,"text":5657},{"id":221,"depth":566,"text":221},{"id":266,"depth":566,"text":266},{"id":405,"depth":566,"text":405},{"id":456,"depth":566,"text":457},{"id":520,"depth":566,"text":520},{"id":538,"depth":566,"text":538},[8697,8698,8699,8700],{"q":6098,"a":6099},{"q":6101,"a":6102},{"q":6104,"a":6105},{"q":6107,"a":6108},[583,6110],{},[1112,6114,591,2440],[8705,8706,8707,8708],{"plan":5630,"price":125,"features":6117,"notes":6118},{"plan":5636,"price":125,"features":6120,"notes":4966},{"plan":5642,"price":6122,"features":6123,"notes":6124},{"plan":728,"price":146,"features":6126,"notes":6127},[6131],{"power":595,"ux":596,"price":596,"cn_support":563,"stability":595},{"title":843,"description":6095},[8713,8714,8715,8716],{"name":6137,"url":6138,"accessed":2498},{"name":6140,"url":6064,"accessed":2498},{"name":6142,"url":6071,"accessed":2498},{"name":6144,"url":6078,"accessed":2498},[610,6148,5528,1113,6149,5745],{"id":6153,"title":2321,"alternatives":8719,"api_compatible":15,"body":8720,"category":577,"chinese_friendly":563,"cover":6801,"description":6802,"domestic":580,"extension":581,"faq":9228,"free":580,"github":15,"languages":9233,"lastVerified":15,"meta":9234,"models":15,"navigation":586,"notSuitable":15,"opensource":586,"path":6034,"pillar":588,"platforms":9235,"priceTable":9236,"pricing":6835,"published":6129,"relatedPlaybooks":9241,"relatedReviews":15,"score":9242,"self_host":586,"seo":9243,"seoTitle":6840,"slug":14,"sources":9244,"stem":6850,"suitable":15,"tagline":6851,"tags":9249,"updated":2498,"verdict":6854,"website":6763,"__hash__":6855},[12,5534,5535],{"type":17,"value":8721,"toc":9216},[8722,8724,8726,8728,8730,8780,8782,8800,8804,8806,8810,8828,8832,8850,8852,8956,8958,8980,8982,8984,9110,9112,9154,9156,9174,9176,9190,9192,9214],[20,8723,23],{"id":22},[25,8725,6161],{},[25,8727,6164],{},[20,8729,33],{"id":33},[35,8731,8732,8736,8740,8744,8748,8752,8756,8760,8764,8768,8772,8776],{},[38,8733,8734,6174],{},[41,8735,6173],{},[38,8737,8738,6180],{},[41,8739,6179],{},[38,8741,8742,6186],{},[41,8743,6185],{},[38,8745,8746,6192],{},[41,8747,6191],{},[38,8749,8750,6198],{},[41,8751,6197],{},[38,8753,8754,6204],{},[41,8755,6203],{},[38,8757,8758,6210],{},[41,8759,6209],{},[38,8761,8762,6216],{},[41,8763,6215],{},[38,8765,8766,6222],{},[41,8767,6221],{},[38,8769,8770,6228],{},[41,8771,6227],{},[38,8773,8774,6234],{},[41,8775,6233],{},[38,8777,8778,6240],{},[41,8779,6239],{},[20,8781,95],{"id":95},[35,8783,8784,8788,8792,8796],{},[38,8785,8786,6250],{},[41,8787,6249],{},[38,8789,8790,6255],{},[41,8791,5110],{},[38,8793,8794,6260],{},[41,8795,5121],{},[38,8797,8798,6265],{},[41,8799,728],{},[151,8801,8802],{},[25,8803,6270],{},[20,8805,6274],{"id":6273},[25,8807,8808],{},[41,8809,167],{},[35,8811,8812,8814,8816,8818,8820,8822,8824,8826],{},[38,8813,6283],{},[38,8815,6286],{},[38,8817,6289],{},[38,8819,6292],{},[38,8821,6295],{},[38,8823,6298],{},[38,8825,6301],{},[38,8827,6304],{},[25,8829,8830],{},[41,8831,195],{},[35,8833,8834,8836,8838,8840,8842,8844,8846,8848],{},[38,8835,6313],{},[38,8837,6316],{},[38,8839,6319],{},[38,8841,6322],{},[38,8843,6325],{},[38,8845,6328],{},[38,8847,6331],{},[38,8849,6334],{},[20,8851,6338],{"id":6337},[3185,8853,8854],{"className":3187,"code":6341,"language":3189,"meta":562,"style":562},[228,8855,8856,8860,8868,8872,8876,8886,8894,8902,8912,8920,8928,8936,8944,8948,8952],{"__ignoreMap":562},[3193,8857,8858],{"class":3195,"line":3196},[3193,8859,6348],{"class":3245},[3193,8861,8862,8864,8866],{"class":3195,"line":566},[3193,8863,6353],{"class":3199},[3193,8865,5769],{"class":3212},[3193,8867,6358],{"class":3203},[3193,8869,8870],{"class":3195,"line":563},[3193,8871,3391],{"emptyLinePlaceholder":586},[3193,8873,8874],{"class":3195,"line":595},[3193,8875,4339],{"class":3245},[3193,8877,8878,8880,8882,8884],{"class":3195,"line":596},[3193,8879,591],{"class":3199},[3193,8881,5748],{"class":3203},[3193,8883,6375],{"class":3212},[3193,8885,6378],{"class":3212},[3193,8887,8888,8890,8892],{"class":3195,"line":3388},[3193,8889,6383],{"class":3212},[3193,8891,6386],{"class":3203},[3193,8893,6378],{"class":3212},[3193,8895,8896,8898,8900],{"class":3195,"line":3394},[3193,8897,6393],{"class":3212},[3193,8899,6396],{"class":3203},[3193,8901,6378],{"class":3212},[3193,8903,8904,8906,8908,8910],{"class":3195,"line":3400},[3193,8905,6403],{"class":3212},[3193,8907,6406],{"class":3203},[3193,8909,6409],{"class":3212},[3193,8911,6378],{"class":3212},[3193,8913,8914,8916,8918],{"class":3195,"line":3406},[3193,8915,6403],{"class":3212},[3193,8917,6418],{"class":3203},[3193,8919,6378],{"class":3212},[3193,8921,8922,8924,8926],{"class":3195,"line":3411},[3193,8923,6403],{"class":3212},[3193,8925,6427],{"class":3203},[3193,8927,6378],{"class":3212},[3193,8929,8930,8932,8934],{"class":3195,"line":3417},[3193,8931,6434],{"class":3212},[3193,8933,6437],{"class":3203},[3193,8935,6378],{"class":3212},[3193,8937,8938,8940,8942],{"class":3195,"line":3422},[3193,8939,6444],{"class":3212},[3193,8941,6447],{"class":3203},[3193,8943,6378],{"class":3212},[3193,8945,8946],{"class":3195,"line":3428},[3193,8947,6454],{"class":3203},[3193,8949,8950],{"class":3195,"line":3434},[3193,8951,3391],{"emptyLinePlaceholder":586},[3193,8953,8954],{"class":3195,"line":4361},[3193,8955,6463],{"class":3245},[25,8957,6466],{},[3185,8959,8960],{"className":3187,"code":6469,"language":3189,"meta":562,"style":562},[228,8961,8962,8968,8976],{"__ignoreMap":562},[3193,8963,8964,8966],{"class":3195,"line":3196},[3193,8965,6476],{"class":3199},[3193,8967,6479],{"class":3203},[3193,8969,8970,8972,8974],{"class":3195,"line":566},[3193,8971,6476],{"class":3199},[3193,8973,6486],{"class":3203},[3193,8975,6489],{"class":3203},[3193,8977,8978],{"class":3195,"line":563},[3193,8979,6494],{"class":3245},[25,8981,6497],{},[20,8983,266],{"id":266},[97,8985,8986,9000],{},[100,8987,8988],{},[103,8989,8990,8992,8994,8996,8998],{},[106,8991,275],{},[106,8993,2321],{},[106,8995,6512],{},[106,8997,6515],{},[106,8999,843],{},[115,9001,9002,9014,9026,9038,9050,9062,9074,9086,9098],{},[103,9003,9004,9006,9008,9010,9012],{},[120,9005,2018],{},[120,9007,6526],{},[120,9009,381],{},[120,9011,381],{},[120,9013,4580],{},[103,9015,9016,9018,9020,9022,9024],{},[120,9017,3291],{},[120,9019,5873],{},[120,9021,381],{},[120,9023,381],{},[120,9025,370],{},[103,9027,9028,9030,9032,9034,9036],{},[120,9029,6549],{},[120,9031,6552],{},[120,9033,6555],{},[120,9035,6558],{},[120,9037,6561],{},[103,9039,9040,9042,9044,9046,9048],{},[120,9041,2328],{},[120,9043,5873],{},[120,9045,5890],{},[120,9047,5890],{},[120,9049,370],{},[103,9051,9052,9054,9056,9058,9060],{},[120,9053,6578],{},[120,9055,6581],{},[120,9057,381],{},[120,9059,381],{},[120,9061,1849],{},[103,9063,9064,9066,9068,9070,9072],{},[120,9065,3799],{},[120,9067,1849],{},[120,9069,381],{},[120,9071,381],{},[120,9073,5890],{},[103,9075,9076,9078,9080,9082,9084],{},[120,9077,6604],{},[120,9079,6607],{},[120,9081,381],{},[120,9083,381],{},[120,9085,370],{},[103,9087,9088,9090,9092,9094,9096],{},[120,9089,6618],{},[120,9091,6621],{},[120,9093,6624],{},[120,9095,6627],{},[120,9097,6621],{},[103,9099,9100,9102,9104,9106,9108],{},[120,9101,5347],{},[120,9103,6636],{},[120,9105,6639],{},[120,9107,6642],{},[120,9109,6645],{},[20,9111,405],{"id":405},[35,9113,9114,9118,9122,9126,9130,9134,9138,9142,9146,9150],{},[38,9115,9116,6655],{},[41,9117,6654],{},[38,9119,9120,6661],{},[41,9121,6660],{},[38,9123,9124,6667],{},[41,9125,6666],{},[38,9127,9128,6673],{},[41,9129,6672],{},[38,9131,9132,6679],{},[41,9133,6678],{},[38,9135,9136,6685],{},[41,9137,6684],{},[38,9139,9140,6691],{},[41,9141,6690],{},[38,9143,9144,6697],{},[41,9145,6696],{},[38,9147,9148,6702],{},[41,9149,6239],{},[38,9151,9152,6708],{},[41,9153,6707],{},[20,9155,457],{"id":456},[35,9157,9158,9160,9162,9164,9166,9168,9170,9172],{},[38,9159,6715],{},[38,9161,6718],{},[38,9163,6721],{},[38,9165,6724],{},[38,9167,6727],{},[38,9169,6730],{},[38,9171,6733],{},[38,9173,6736],{},[20,9175,520],{"id":520},[35,9177,9178,9182,9186],{},[38,9179,9180],{},[524,9181,6745],{"href":6112},[38,9183,9184],{},[524,9185,6041],{"href":6040},[38,9187,9188],{},[524,9189,6047],{"href":6046},[20,9191,538],{"id":538},[223,9193,9194,9199,9204,9209],{},[38,9195,6760,9196],{},[524,9197,6763],{"href":6763,"rel":9198},[552],[38,9200,6767,9201],{},[524,9202,6770],{"href":6770,"rel":9203},[552],[38,9205,6774,9206],{},[524,9207,6777],{"href":6777,"rel":9208},[552],[38,9210,6781,9211],{},[524,9212,6784],{"href":6784,"rel":9213},[552],[3864,9215,6788],{},{"title":562,"searchDepth":563,"depth":563,"links":9217},[9218,9219,9220,9221,9222,9223,9224,9225,9226,9227],{"id":22,"depth":566,"text":23},{"id":33,"depth":566,"text":33},{"id":95,"depth":566,"text":95},{"id":6273,"depth":566,"text":6274},{"id":6337,"depth":566,"text":6338},{"id":266,"depth":566,"text":266},{"id":405,"depth":566,"text":405},{"id":456,"depth":566,"text":457},{"id":520,"depth":566,"text":520},{"id":538,"depth":566,"text":538},[9229,9230,9231,9232],{"q":6805,"a":6806},{"q":6808,"a":6809},{"q":6811,"a":6812},{"q":6814,"a":6815},[583,6110],{},[1112,6114,591,2440],[9237,9238,9239,9240],{"plan":6821,"price":125,"features":6822,"notes":6823},{"plan":5110,"price":6825,"features":6826,"notes":6827},{"plan":5121,"price":6829,"features":6830,"notes":6831},{"plan":728,"price":146,"features":6833,"notes":6834},[6837],{"power":596,"ux":595,"price":596,"cn_support":563,"stability":596},{"title":2321,"description":6802},[9245,9246,9247,9248],{"name":6843,"url":6763,"accessed":2498},{"name":6845,"url":6770,"accessed":2498},{"name":6847,"url":6777,"accessed":2498},{"name":6849,"url":6784,"accessed":2498},[610,2494,6853,5528,3934,1112,2321],1785428435745]