[{"data":1,"prerenderedAt":1095},["ShallowReactive",2],{"header-counts":3,"footer-counts":6,"compare-anythingllm-vs-ragflow":8,"compare-a-anythingllm":9,"compare-b-ragflow":583},{"tools":4,"reviews":5},98,27,{"tools":4,"reviews":5,"playbooks":7,"news":5},22,null,{"id":10,"title":11,"alternatives":12,"api_compatible":8,"body":16,"category":545,"chinese_friendly":531,"cover":546,"description":547,"domestic":548,"extension":549,"faq":8,"free":548,"github":526,"languages":550,"lastVerified":552,"meta":553,"models":8,"navigation":554,"notSuitable":8,"opensource":554,"path":555,"pillar":556,"platforms":557,"priceTable":8,"pricing":562,"published":563,"relatedPlaybooks":8,"relatedReviews":8,"score":564,"self_host":548,"seo":567,"seoTitle":568,"slug":569,"sources":570,"stem":573,"suitable":8,"tagline":574,"tags":575,"updated":552,"verdict":581,"website":518,"__hash__":582},"tools\u002Ftools\u002Fagent\u002Fplatform\u002Fanythingllm.md","AnythingLLM",[13,14,15],"agent\u002Fplatform\u002Fdify","agent\u002Fplatform\u002Ffastgpt","agent\u002Fplatform\u002Flangflow",{"type":17,"value":18,"toc":529},"minimark",[19,24,28,31,34,93,96,151,157,161,169,194,199,219,222,251,254,380,383,424,428,457,461,467,473,479,485,488,504,507,512],[20,21,23],"h2",{"id":22},"tldr","TL;DR",[25,26,27],"p",{},"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。",[25,29,30],{},"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。",[20,32,33],{"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",{},"私有化部署","：Docker \u002F 桌面应用（Win\u002FMac\u002FLinux），数据完全在内网",[38,46,47,50],{},[41,48,49],{},"Workspaces 知识库隔离","：不同工作区独立向量库 + 文档 + 对话历史",[38,52,53,56],{},[41,54,55],{},"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",[38,58,59,62],{},[41,60,61],{},"任意模型接入","：OpenAI \u002F Claude \u002F Azure \u002F Ollama \u002F LM Studio \u002F 本地模型",[38,64,65,68],{},[41,66,67],{},"多向量数据库","：内置 LanceDB，可选 Chroma \u002F Pinecone \u002F Weaviate \u002F Qdrant",[38,70,71,74],{},[41,72,73],{},"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",[38,76,77,80],{},[41,78,79],{},"Agent 能力","：内置 Web 搜索 \u002F RAG 搜索 \u002F SQL 查询等工具调用",[38,82,83,86],{},[41,84,85],{},"嵌入向量","：支持自定义 embedding 模型，兼容 OpenAI \u002F 本地嵌入",[38,88,89,92],{},[41,90,91],{},"API 接口","：提供完整 REST API，可集成到外部系统",[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",[115,116,117,129,140],"tbody",{},[103,118,119,123,126],{},[120,121,122],"td",{},"开源版",[120,124,125],{},"$0",[120,127,128],{},"完整功能，MIT 协议，自托管",[103,130,131,134,137],{},[120,132,133],{},"Cloud",[120,135,136],{},"$30\u002F月起",[120,138,139],{},"托管服务，免去运维，含团队协作",[103,141,142,145,148],{},[120,143,144],{},"Enterprise",[120,146,147],{},"联系销售",[120,149,150],{},"SSO \u002F 审计日志 \u002F 私有部署支持",[152,153,154],"blockquote",{},[25,155,156],{},"价格信息基于 2026-07 官网，可能调整。",[20,158,160],{"id":159},"体验与评测资料整理","体验与评测（资料整理）",[152,162,163],{},[25,164,165,166],{},"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",[41,167,168],{},"亮点：",[35,170,171,179,182,185,188,191],{},[38,172,173,174,178],{},"Docker 部署极快，一条 ",[175,176,177],"code",{},"docker-compose up"," 起来就能用",[38,180,181],{},"Workspaces 隔离设计实用，不同部门知识库互不干扰",[38,183,184],{},"接 Ollama 本地模型完全离线运行，数据不出内网",[38,186,187],{},"桌面应用适合个人用户，安装即用零配置",[38,189,190],{},"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",[38,192,193],{},"多用户权限管理是开源 RAG 平台中少有的完整实现",[25,195,196],{},[41,197,198],{},"踩坑：",[35,200,201,204,207,210,213,216],{},[38,202,203],{},"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",[38,205,206],{},"大文件（100MB+ PDF）处理偶尔超时，需调超时参数",[38,208,209],{},"Agent 能力有限，复杂工具链编排不如 Dify",[38,211,212],{},"向量库默认 LanceDB 在数据量大时查询变慢，建议切 Qdrant \u002F Chroma",[38,214,215],{},"UI 偶有卡顿，文档列表加载慢",[38,217,218],{},"中文文档的 OCR 需要额外配置，默认对扫描件支持有限",[20,220,221],{"id":221},"上手",[223,224,225,232,239,242,245,248],"ol",{},[38,226,227,228,231],{},"Docker 部署：",[175,229,230],{},"docker-compose up -d","（官方提供 docker-compose.yml）",[38,233,234,235,238],{},"首次访问 ",[175,236,237],{},"http:\u002F\u002Flocalhost:3001","，创建管理员账号",[38,240,241],{},"Settings → LLM Provider 配置模型（OpenAI API Key 或 Ollama 地址）",[38,243,244],{},"创建 Workspace → 上传文档（PDF\u002FWord\u002FTXT）",[38,246,247],{},"等待文档向量化完成，在 Chat 中开始问答",[38,249,250],{},"Settings → Users 添加团队成员并分配工作区权限",[20,252,253],{"id":253},"对比",[97,255,256,274],{},[100,257,258],{},[103,259,260,263,265,268,271],{},[106,261,262],{},"维度",[106,264,11],{},[106,266,267],{},"Dify",[106,269,270],{},"FastGPT",[106,272,273],{},"Langflow",[115,275,276,291,307,321,337,351,364],{},[103,277,278,281,284,287,289],{},[120,279,280],{},"部署门槛",[120,282,283],{},"极低",[120,285,286],{},"中",[120,288,286],{},[120,290,286],{},[103,292,293,296,299,302,304],{},[120,294,295],{},"多用户权限",[120,297,298],{},"✅ 完整",[120,300,301],{},"✅",[120,303,301],{},[120,305,306],{},"❌",[103,308,309,312,314,317,319],{},[120,310,311],{},"RAG 精度",[120,313,286],{},[120,315,316],{},"高",[120,318,316],{},[120,320,286],{},[103,322,323,326,329,332,334],{},[120,324,325],{},"Agent 编排",[120,327,328],{},"基础",[120,330,331],{},"强",[120,333,286],{},[120,335,336],{},"强（可视化）",[103,338,339,342,345,347,349],{},[120,340,341],{},"模型接入",[120,343,344],{},"丰富",[120,346,344],{},[120,348,344],{},[120,350,344],{},[103,352,353,356,358,360,362],{},[120,354,355],{},"桌面应用",[120,357,301],{},[120,359,306],{},[120,361,306],{},[120,363,306],{},[103,365,366,369,372,375,378],{},[120,367,368],{},"协议",[120,370,371],{},"MIT",[120,373,374],{},"Apache 2.0",[120,376,377],{},"FastGPT Open",[120,379,371],{},[20,381,382],{"id":382},"避坑",[35,384,385,391,397,403,412,418],{},[38,386,387,390],{},[41,388,389],{},"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",[38,392,393,396],{},[41,394,395],{},"LanceDB 大数据量变慢","：文档超过 1 万条建议切 Qdrant 或 Chroma",[38,398,399,402],{},[41,400,401],{},"大文件超时","：调整 Docker 超时配置，或拆分文档上传",[38,404,405,408,409],{},[41,406,407],{},"Ollama 连接","：Docker 内访问宿主机 Ollama 需用 ",[175,410,411],{},"host.docker.internal",[38,413,414,417],{},[41,415,416],{},"embedding 模型选择","：中文场景建议用 bge-large-zh 而非默认 OpenAI embedding",[38,419,420,423],{},[41,421,422],{},"不要当生产级 Agent 平台用","：Agent 能力是辅助，复杂编排上 Dify",[20,425,427],{"id":426},"适合-不适合","适合 \u002F 不适合",[35,429,430,433,436,439,442,445,448,451,454],{},[38,431,432],{},"✅ 企业内网私有化 AI 知识库",[38,434,435],{},"✅ 团队共享 AI 助手 + 多用户权限管理",[38,437,438],{},"✅ 接 Ollama 完全离线运行",[38,440,441],{},"✅ 个人桌面端快速体验 RAG",[38,443,444],{},"✅ 需要快速验证 RAG 概念的原型项目",[38,446,447],{},"❌ 需要复杂 Agent 工作流编排（用 Dify \u002F Langflow）",[38,449,450],{},"❌ 需要极致 RAG 召回精度（用 RAGFlow \u002F FastGPT）",[38,452,453],{},"❌ 大规模并发生产服务（架构未做高可用）",[38,455,456],{},"❌ 需要深度文档解析（复杂表格\u002F公式\u002F扫描件）",[20,458,460],{"id":459},"faq","FAQ",[25,462,463,466],{},[41,464,465],{},"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。",[25,468,469,472],{},[41,470,471],{},"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。",[25,474,475,478],{},[41,476,477],{},"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。",[25,480,481,484],{},[41,482,483],{},"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。",[20,486,487],{"id":487},"相关阅读",[25,489,490,495,496,495,500],{},[491,492,494],"a",{"href":493},"\u002Fagent\u002Fplatform\u002Fragflow.html","RAGFlow"," · ",[491,497,499],{"href":498},"\u002Fagent\u002Fplatform\u002Fflowise.html","Flowise",[491,501,503],{"href":502},"\u002Fagent\u002Fgeneral\u002Fperplexity.html","Perplexity",[20,505,506],{"id":506},"来源",[152,508,509],{},[25,510,511],{},"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。",[35,513,514,522],{},[38,515,516],{},[491,517,521],{"href":518,"rel":519},"https:\u002F\u002Fuseanything.com",[520],"nofollow","官网",[38,523,524],{},[491,525,528],{"href":526,"rel":527},"https:\u002F\u002Fgithub.com\u002FMintplex-Labs\u002Fanything-llm",[520],"GitHub",{"title":530,"searchDepth":531,"depth":531,"links":532},"",3,[533,535,536,537,538,539,540,541,542,543,544],{"id":22,"depth":534,"text":23},2,{"id":33,"depth":534,"text":33},{"id":95,"depth":534,"text":95},{"id":159,"depth":534,"text":160},{"id":221,"depth":534,"text":221},{"id":253,"depth":534,"text":253},{"id":382,"depth":534,"text":382},{"id":426,"depth":534,"text":427},{"id":459,"depth":534,"text":460},{"id":487,"depth":534,"text":487},{"id":506,"depth":534,"text":506},"platform","\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。",false,"md",[551],"en","2026-07-30",{},true,"\u002Ftools\u002Fagent\u002Fplatform\u002Fanythingllm","agent",[558,559,560,561],"windows","macos","linux","docker","Free \u002F 开源（MIT）\u002F Cloud","2026-07-05",{"power":531,"ux":565,"price":566,"cn_support":531,"stability":531},4,5,{"title":11,"description":547},"AnythingLLM - 开源私有部署 LLM 平台评测 | AIHO","agent\u002Fplatform\u002Fanythingllm",[571,572],{"title":521,"url":518},{"title":528,"url":526},"tools\u002Fagent\u002Fplatform\u002Fanythingllm","开源私有部署 LLM 平台，一站式 RAG + Agent + 多用户",[576,577,578,579,580],"agent-platform","opensource","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":584,"title":494,"alternatives":585,"api_compatible":8,"body":586,"category":545,"chinese_friendly":565,"cover":1074,"description":1075,"domestic":548,"extension":549,"faq":8,"free":548,"github":1059,"languages":1076,"lastVerified":552,"meta":1078,"models":8,"navigation":554,"notSuitable":8,"opensource":554,"path":1079,"pillar":556,"platforms":1080,"priceTable":8,"pricing":1081,"published":563,"relatedPlaybooks":8,"relatedReviews":8,"score":1082,"self_host":548,"seo":1083,"seoTitle":1084,"slug":1085,"sources":1086,"stem":1089,"suitable":8,"tagline":1090,"tags":1091,"updated":552,"verdict":1093,"website":1053,"__hash__":1094},"tools\u002Ftools\u002Fagent\u002Fplatform\u002Fragflow.md",[13,14,569],{"type":17,"value":587,"toc":1061},[588,590,593,596,598,652,654,696,700,702,708,728,732,755,757,787,789,918,920,968,970,1002,1004,1010,1016,1022,1028,1030,1041,1043,1047],[20,589,23],{"id":22},[25,591,592],{},"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）。",[25,594,595],{},"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。",[20,597,33],{"id":33},[35,599,600,606,612,618,624,630,636,641,647],{},[38,601,602,605],{},[41,603,604],{},"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",[38,607,608,611],{},[41,609,610],{},"OCR 引擎","：内置 PaddleOCR \u002F DeepDOC，支持中英文扫描件识别",[38,613,614,617],{},[41,615,616],{},"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",[38,619,620,623],{},[41,621,622],{},"高召回率","：混合检索（全文 + 向量）+ 重排序（Rerank），召回精度高",[38,625,626,629],{},[41,627,628],{},"引用溯源","：回答标注来源文档 + 页码 + 原文片段，可验证",[38,631,632,635],{},[41,633,634],{},"多模型接入","：OpenAI \u002F Claude \u002F Ollama \u002F 通义千问 \u002F 智谱 \u002F 月之暗面",[38,637,638,640],{},[41,639,67],{},"：Elasticsearch \u002F Infinity（自研）\u002F Chroma",[38,642,643,646],{},[41,644,645],{},"知识库管理","：多知识库 + 文档分类 + 解析状态监控",[38,648,649,651],{},[41,650,91],{},"：完整 REST API + SDK，可集成到外部系统",[20,653,95],{"id":95},[97,655,656,666],{},[100,657,658],{},[103,659,660,662,664],{},[106,661,108],{},[106,663,95],{},[106,665,113],{},[115,667,668,677,687],{},[103,669,670,672,674],{},[120,671,122],{},[120,673,125],{},[120,675,676],{},"完整功能，Apache 2.0，自托管",[103,678,679,681,684],{},[120,680,133],{},[120,682,683],{},"按量付费",[120,685,686],{},"托管服务，免运维",[103,688,689,691,693],{},[120,690,144],{},[120,692,147],{},[120,694,695],{},"私有部署 + 技术支持 + 定制",[152,697,698],{},[25,699,156],{},[20,701,160],{"id":159},[152,703,704],{},[25,705,165,706],{},[41,707,168],{},[35,709,710,713,716,719,722,725],{},[38,711,712],{},"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",[38,714,715],{},"扫描件 OCR 效果好，中文印刷体识别准确率高",[38,717,718],{},"引用溯源到页码 + 原文片段，回答可信度高",[38,720,721],{},"混合检索 + Rerank 召回精度明显优于纯向量检索",[38,723,724],{},"中国团队出品，中文文档和社区支持好，Issue 响应快",[38,726,727],{},"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",[25,729,730],{},[41,731,198],{},[35,733,734,737,740,743,746,749,752],{},[38,735,736],{},"资源消耗大——Elasticsearch + Redis + MinIO + RAGFlow 本身，至少 16GB 内存",[38,738,739],{},"部署较重，Docker Compose 起来 5+ 容器，配置复杂",[38,741,742],{},"大文件解析慢——100 页 PDF 解析 + 切片可能 5-10 分钟",[38,744,745],{},"UI 仍有粗糙处，文档管理界面交互不够流畅",[38,747,748],{},"Agent 能力弱——RAG 问答是强项，复杂工具调用 \u002F 多步推理不如 Dify",[38,750,751],{},"解析失败的重试机制不完善，偶尔卡在 parsing 状态",[38,753,754],{},"版本迭代快，升级需注意数据迁移",[20,756,221],{"id":221},[223,758,759,762,768,774,781,784],{},[38,760,761],{},"系统准备：确保 16GB+ 内存 + Docker + Docker Compose",[38,763,764,765],{},"克隆仓库：",[175,766,767],{},"git clone https:\u002F\u002Fgithub.com\u002Finfiniflow\u002Fragflow.git",[38,769,770,771],{},"启动服务：",[175,772,773],{},"cd ragflow\u002Fdocker && docker compose up -d",[38,775,776,777,780],{},"访问 ",[175,778,779],{},"http:\u002F\u002Flocalhost:80","，注册管理员账号",[38,782,783],{},"配置模型：Settings → Model Providers 添加 LLM + Embedding + Rerank",[38,785,786],{},"创建知识库 → 上传文档 → 等待解析完成 → 开始问答",[20,788,253],{"id":253},[97,790,791,805],{},[100,792,793],{},[103,794,795,797,799,801,803],{},[106,796,262],{},[106,798,494],{},[106,800,267],{},[106,802,270],{},[106,804,11],{},[115,806,807,822,835,850,863,876,888,902],{},[103,808,809,812,815,817,819],{},[120,810,811],{},"文档解析",[120,813,814],{},"✅ 最强",[120,816,286],{},[120,818,331],{},[120,820,821],{},"弱",[103,823,824,827,829,831,833],{},[120,825,826],{},"表格识别",[120,828,301],{},[120,830,306],{},[120,832,301],{},[120,834,306],{},[103,836,837,840,843,845,848],{},[120,838,839],{},"OCR",[120,841,842],{},"✅ 内置",[120,844,306],{},[120,846,847],{},"需配置",[120,849,847],{},[103,851,852,855,857,859,861],{},[120,853,854],{},"召回精度",[120,856,316],{},[120,858,316],{},[120,860,316],{},[120,862,286],{},[103,864,865,867,870,872,874],{},[120,866,628],{},[120,868,869],{},"✅ 页码+片段",[120,871,301],{},[120,873,301],{},[120,875,301],{},[103,877,878,880,882,884,886],{},[120,879,79],{},[120,881,821],{},[120,883,331],{},[120,885,286],{},[120,887,328],{},[103,889,890,893,895,897,899],{},[120,891,892],{},"资源消耗",[120,894,316],{},[120,896,286],{},[120,898,286],{},[120,900,901],{},"低",[103,903,904,907,910,913,915],{},[120,905,906],{},"中文支持",[120,908,909],{},"✅ 优秀",[120,911,912],{},"好",[120,914,912],{},[120,916,917],{},"一般",[20,919,382],{"id":382},[35,921,922,928,938,944,950,956,962],{},[38,923,924,927],{},[41,925,926],{},"资源一定要够","：低于 16GB 内存别部署，ES + Redis + MinIO 都吃内存",[38,929,930,933,934,937],{},[41,931,932],{},"Elasticsearch 配置","：默认 JVM 堆偏小，大知识库调 ",[175,935,936],{},"ES_JAVA_OPTS"," 到 4-8GB",[38,939,940,943],{},[41,941,942],{},"大文件拆分上传","：超过 100 页的 PDF 拆成小文件，解析更稳定",[38,945,946,949],{},[41,947,948],{},"解析失败检查格式","：加密 PDF \u002F 损坏文件会卡住，上传前检查",[38,951,952,955],{},[41,953,954],{},"Rerank 模型别省","：召回精度提升的关键，用 bge-reranker 或 Cohere Rerank",[38,957,958,961],{},[41,959,960],{},"不要当 Agent 平台用","：RAG 问答是核心，复杂工具调用上 Dify",[38,963,964,967],{},[41,965,966],{},"定期备份","：ES 数据 + MinIO 文件，升级前完整快照",[20,969,427],{"id":426},[35,971,972,975,978,981,984,987,990,993,996,999],{},[38,973,974],{},"✅ 需要精准文档问答的企业知识库",[38,976,977],{},"✅ 复杂文档（表格 \u002F 图文 \u002F 扫描件）RAG 场景",[38,979,980],{},"✅ 中文 RAG 需求（国产模型 + 中文 OCR）",[38,982,983],{},"✅ 对召回率和引用溯源要求高的业务",[38,985,986],{},"✅ 有运维能力的团队私有化部署",[38,988,989],{},"❌ 需要复杂 Agent 编排（用 Dify）",[38,991,992],{},"❌ 资源有限的小服务器（至少 16GB 内存）",[38,994,995],{},"❌ 需要精美 C 端 UI 的产品",[38,997,998],{},"❌ 无运维能力的团队（用 Cloud 版或 FastGPT）",[38,1000,1001],{},"❌ 纯英文简单文档场景（AnythingLLM 更轻量）",[20,1003,460],{"id":459},[25,1005,1006,1009],{},[41,1007,1008],{},"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。",[25,1011,1012,1015],{},[41,1013,1014],{},"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。",[25,1017,1018,1021],{},[41,1019,1020],{},"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。",[25,1023,1024,1027],{},[41,1025,1026],{},"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。",[20,1029,487],{"id":487},[25,1031,1032,495,1035,495,1037],{},[491,1033,11],{"href":1034},"\u002Fagent\u002Fplatform\u002Fanythingllm.html",[491,1036,499],{"href":498},[491,1038,1040],{"href":1039},"\u002Fcoding\u002Fapi\u002Flangfuse.html","Langfuse",[20,1042,506],{"id":506},[152,1044,1045],{},[25,1046,511],{},[35,1048,1049,1055],{},[38,1050,1051],{},[491,1052,521],{"href":1053,"rel":1054},"https:\u002F\u002Fragflow.io",[520],[38,1056,1057],{},[491,1058,528],{"href":1059,"rel":1060},"https:\u002F\u002Fgithub.com\u002Finfiniflow\u002Fragflow",[520],{"title":530,"searchDepth":531,"depth":531,"links":1062},[1063,1064,1065,1066,1067,1068,1069,1070,1071,1072,1073],{"id":22,"depth":534,"text":23},{"id":33,"depth":534,"text":33},{"id":95,"depth":534,"text":95},{"id":159,"depth":534,"text":160},{"id":221,"depth":534,"text":221},{"id":253,"depth":534,"text":253},{"id":382,"depth":534,"text":382},{"id":426,"depth":534,"text":427},{"id":459,"depth":534,"text":460},{"id":487,"depth":534,"text":487},{"id":506,"depth":534,"text":506},"\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。",[551,1077],"zh",{},"\u002Ftools\u002Fagent\u002Fplatform\u002Fragflow",[560,561],"Free \u002F 开源（Apache 2.0）\u002F Cloud",{"power":565,"ux":531,"price":566,"cn_support":565,"stability":531},{"title":494,"description":1075},"RAGFlow - 开源 RAG 引擎评测与部署 | AIHO","agent\u002Fplatform\u002Fragflow",[1087,1088],{"title":521,"url":1053},{"title":528,"url":1059},"tools\u002Fagent\u002Fplatform\u002Fragflow","开源 RAG 引擎，深度文档解析 + 高召回率",[576,577,579,1092,578],"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",1785428441240]