[{"data":1,"prerenderedAt":1070},["ShallowReactive",2],{"header-counts":3,"footer-counts":6,"compare-lm-studio-vs-lobe-chat":9,"compare-a-lm-studio":10,"compare-b-lobe-chat":561},{"tools":4,"reviews":5},78,26,{"tools":4,"reviews":5,"playbooks":7,"news":8},22,21,null,{"id":11,"title":12,"alternatives":13,"api_compatible":9,"body":18,"category":493,"chinese_friendly":480,"cover":494,"description":495,"domestic":496,"extension":497,"faq":498,"free":496,"github":9,"languages":511,"lastVerified":9,"meta":514,"models":9,"navigation":515,"notSuitable":9,"opensource":496,"path":516,"pillar":517,"platforms":518,"priceTable":522,"pricing":531,"published":532,"relatedPlaybooks":533,"relatedReviews":9,"score":536,"self_host":515,"seo":539,"seoTitle":9,"slug":540,"sources":541,"stem":549,"suitable":9,"tagline":550,"tags":551,"updated":544,"verdict":558,"website":559,"__hash__":560},"tools\u002Ftools\u002Fcoding\u002Flocal\u002Flm-studio.md","LM Studio",[14,15,16,17],"coding\u002Flocal\u002Follama","coding\u002Flocal\u002Fopen-webui","coding\u002Flocal\u002Fcherry-studio","coding\u002Flocal\u002Flobe-chat",{"type":19,"value":20,"toc":478},"minimark",[21,26,35,38,41,101,104,118,124,128,133,150,155,172,175,202,205,359,362,394,398,421,424,451,454],[22,23,25],"h2",{"id":24},"tldr","TL;DR",[27,28,29,30,34],"p",{},"LM Studio 是 Windows \u002F macOS \u002F Linux 桌面应用，让你像浏览 App Store 一样发现、下载、运行本地大模型（GGUF \u002F MLX 格式）。底层基于 llama.cpp + MLX，Mac M 系列原生优化。0.3+ 起新增 Headless 模式 + ",[31,32,33],"code",{},"lms"," CLI，可在服务器跑 OpenAI 兼容 API（默认 :1234）。个人 \u002F 评估完全免费，商用咨询。",[27,36,37],{},"适合：本地 LLM 入门 \u002F 评估、Mac 用户、需要 GUI 调参 \u002F 模型比较、想给 IDE \u002F 应用接本地 OpenAI 兼容 endpoint 的开发者。不适合：多用户并发生产服务（用 vLLM）、嵌入式 \u002F 边缘部署（用 llama.cpp）、纯 CLI 工作流（用 Ollama）。",[22,39,40],{"id":40},"核心能力",[42,43,44,52,58,64,74,83,89,95],"ul",{},[45,46,47,51],"li",{},[48,49,50],"strong",{},"模型浏览器","：内置 Hugging Face 检索，按 GGUF \u002F MLX \u002F 大小筛选、一键下载",[45,53,54,57],{},[48,55,56],{},"聊天界面","：System Prompt \u002F temperature \u002F top-p \u002F context size 可视化调参",[45,59,60,63],{},[48,61,62],{},"多模型并存 \u002F 切换","：同时加载多模型在不同会话中比较",[45,65,66,69,70,73],{},[48,67,68],{},"OpenAI 兼容 Local Server","：",[31,71,72],{},"http:\u002F\u002Flocalhost:1234\u002Fv1","，任何 SDK 即接即用",[45,75,76,69,79,82],{},[48,77,78],{},"Headless \u002F CLI",[31,80,81],{},"lms server start --port 1234","，无 GUI 可跑",[45,84,85,88],{},[48,86,87],{},"PDF \u002F 文档对话","：内置基础 RAG，丢文件就能聊",[45,90,91,94],{},[48,92,93],{},"MLX 原生支持（Mac）","：M1+ 上比 GGUF + Metal 快 30–50%",[45,96,97,100],{},[48,98,99],{},"持续批处理","：Codersera 2026 测得 50–90 tok\u002Fs（消费级 GPU + 中等模型）",[22,102,103],{"id":103},"价格",[42,105,106,112],{},[45,107,108,111],{},[48,109,110],{},"个人 \u002F 评估","：免费，全功能可用",[45,113,114,117],{},[48,115,116],{},"商用","：邮件 \u002F 官网联系 LM Studio 团队",[119,120,121],"blockquote",{},[27,122,123],{},"模型本身免费（开源权重），LM Studio 不抽水任何 token 费用。",[22,125,127],{"id":126},"实测mac-m2-pro-qwen3-coder-7b-gguf-q4_k_m","实测（Mac M2 Pro + Qwen3-Coder-7B GGUF Q4_K_M）",[27,129,130],{},[48,131,132],{},"亮点：",[42,134,135,138,141,144,147],{},[45,136,137],{},"模型浏览器极舒服：搜「qwen3-coder」直接列出 GGUF + MLX 各 quant，标硬件兼容度",[45,139,140],{},"加载 7B Q4 模型 \u003C 3 秒，生成 ~75 tok\u002Fs",[45,142,143],{},"Local Server 开了 Cursor 直接接 baseURL → 本地代码补全零成本",[45,145,146],{},"MLX 版同模型 ~110 tok\u002Fs，差距显著",[45,148,149],{},"多窗口加载 2 个模型并排测，调 prompt 直观",[27,151,152],{},[48,153,154],{},"踩坑：",[42,156,157,160,163,166,169],{},[45,158,159],{},"模型库依赖 Hugging Face，国内访问要镜像 \u002F 代理",[45,161,162],{},"GPU 显存吃满后会自动 offload 到 CPU，无提示就慢下来",[45,164,165],{},"Headless 模式相对 Ollama 偏新，文档稍少",[45,167,168],{},"闭源应用（虽免费），不适合企业合规挂钩",[45,170,171],{},"中文 UI 可用但部分菜单仍英文",[22,173,174],{"id":174},"上手",[176,177,178,181,184,187,190,197],"ol",{},[45,179,180],{},"lmstudio.ai 下载（Mac \u002F Windows \u002F Linux）",[45,182,183],{},"打开 → Discover 标签 → 搜模型（如 qwen3-coder、deepseek-v3 GGUF\u002FMLX）→ Download",[45,185,186],{},"Chat 标签 → 选模型 → 调参聊天",[45,188,189],{},"Local Server 标签 → Start Server → 默认端口 1234",[45,191,192,193,196],{},"在你的应用里：",[31,194,195],{},"baseURL = \"http:\u002F\u002Flocalhost:1234\u002Fv1\"","，API Key 任意",[45,198,199,200],{},"Headless：",[31,201,81],{},[22,203,204],{"id":204},"对比",[206,207,208,229],"table",{},[209,210,211],"thead",{},[212,213,214,218,220,223,226],"tr",{},[215,216,217],"th",{},"维度",[215,219,12],{},[215,221,222],{},"Ollama",[215,224,225],{},"Open WebUI",[215,227,228],{},"llama.cpp",[230,231,232,250,267,283,298,313,328,343],"tbody",{},[212,233,234,238,241,244,247],{},[235,236,237],"td",{},"形态",[235,239,240],{},"GUI + CLI",[235,242,243],{},"CLI Daemon",[235,245,246],{},"Docker UI",[235,248,249],{},"二进制",[212,251,252,255,258,261,264],{},[235,253,254],{},"模型浏览",[235,256,257],{},"✅ 内置",[235,259,260],{},"CLI pull",[235,262,263],{},"无",[235,265,266],{},"手动",[212,268,269,272,275,278,281],{},[235,270,271],{},"参数调优 GUI",[235,273,274],{},"✅",[235,276,277],{},"❌",[235,279,280],{},"部分",[235,282,277],{},[212,284,285,288,291,294,296],{},[235,286,287],{},"OpenAI 兼容 API",[235,289,290],{},"✅ :1234",[235,292,293],{},"✅ :11434",[235,295,274],{},[235,297,274],{},[212,299,300,303,305,308,311],{},[235,301,302],{},"MLX (Mac)",[235,304,274],{},[235,306,307],{},"✅ 0.19+",[235,309,310],{},"–",[235,312,310],{},[212,314,315,318,321,323,325],{},[235,316,317],{},"多用户并发",[235,319,320],{},"弱",[235,322,320],{},[235,324,274],{},[235,326,327],{},"中",[212,329,330,333,336,339,341],{},[235,331,332],{},"开源",[235,334,335],{},"闭源（免费）",[235,337,338],{},"MIT",[235,340,338],{},[235,342,338],{},[212,344,345,348,351,354,356],{},[235,346,347],{},"上手难度",[235,349,350],{},"极低",[235,352,353],{},"低",[235,355,327],{},[235,357,358],{},"高",[22,360,361],{"id":361},"避坑",[42,363,364,370,376,382,388],{},[45,365,366,369],{},[48,367,368],{},"国内下模型走镜像","：HF 直连慢 \u002F 卡，配 HF_ENDPOINT=hf-mirror.com",[45,371,372,375],{},[48,373,374],{},"显存爆 ≠ 报错","：GPU 装不下会无声 offload 到 CPU，关注生成速度，必要时降 quant 或换小模型",[45,377,378,381],{},[48,379,380],{},"MLX 优先（Mac M 系列）","：能下 MLX 版就别下 GGUF，速度差距明显",[45,383,384,387],{},[48,385,386],{},"Local Server 暴露要谨慎","：默认 0.0.0.0 + 无鉴权，对外开放前加反代 + Bearer",[45,389,390,393],{},[48,391,392],{},"闭源合规要核","：企业内部使用前查 license；商用必须联系官方",[22,395,397],{"id":396},"适合-不适合","适合 \u002F 不适合",[42,399,400,403,406,409,412,415,418],{},[45,401,402],{},"✅ 本地 LLM 入门 \u002F 评估",[45,404,405],{},"✅ Mac M 系列用户",[45,407,408],{},"✅ 想给 Cursor \u002F Cline 接本地 OpenAI 兼容 endpoint",[45,410,411],{},"✅ 需要 GUI 调参 \u002F 模型比较",[45,413,414],{},"❌ 多用户并发生产服务",[45,416,417],{},"❌ 嵌入式 \u002F 边缘设备",[45,419,420],{},"❌ 强合规 \u002F 必须开源审计",[22,422,423],{"id":423},"相关阅读",[42,425,426,433,439,445],{},[45,427,428],{},[429,430,432],"a",{"href":431},"\u002Ftools\u002Fcoding\u002Flocal\u002Follama","Ollama 评测",[45,434,435],{},[429,436,438],{"href":437},"\u002Ftools\u002Fcoding\u002Flocal\u002Fopen-webui","Open WebUI 评测",[45,440,441],{},[429,442,444],{"href":443},"\u002Ftools\u002Fcoding\u002Flocal\u002Fcherry-studio","Cherry Studio 评测",[45,446,447],{},[429,448,450],{"href":449},"\u002Fplaybook\u002Fonboarding\u002Fclaude-code-getting-started","Claude Code 上手 Playbook",[22,452,453],{"id":453},"来源",[176,455,456,464,471],{},[45,457,458,459],{},"LM Studio 官网 ",[429,460,461],{"href":461,"rel":462},"https:\u002F\u002Flmstudio.ai\u002F",[463],"nofollow",[45,465,466,467],{},"Codersera — LM Studio Complete Guide 2026 ",[429,468,469],{"href":469,"rel":470},"https:\u002F\u002Fcodersera.com\u002Fblog\u002Flm-studio-complete-guide-2026\u002F",[463],[45,472,473,474],{},"Codersera — Ollama vs LM Studio vs vLLM vs llama.cpp vs MLX 2026 ",[429,475,476],{"href":476,"rel":477},"https:\u002F\u002Fcodersera.com\u002Fblog\u002Follama-vs-lm-studio-vs-vllm-vs-llama-cpp-vs-mlx-2026\u002F",[463],{"title":479,"searchDepth":480,"depth":480,"links":481},"",3,[482,484,485,486,487,488,489,490,491,492],{"id":24,"depth":483,"text":25},2,{"id":40,"depth":483,"text":40},{"id":103,"depth":483,"text":103},{"id":126,"depth":483,"text":127},{"id":174,"depth":483,"text":174},{"id":204,"depth":483,"text":204},{"id":361,"depth":483,"text":361},{"id":396,"depth":483,"text":397},{"id":423,"depth":483,"text":423},{"id":453,"depth":483,"text":453},"local","\u002Fimg\u002Ftools\u002Flm-studio.webp","LM Studio 真实评测：跨平台桌面应用，运行本地 GGUF \u002F MLX 大模型。50–90 tok\u002Fs 持续批处理、OpenAI 兼容本地 API（默认端口 1234）、Headless 模式、Mac \u002F Win 双端。对个人开发者免费，企业咨询。",false,"md",[499,502,505,508],{"q":500,"a":501},"和 Ollama 怎么选？","LM Studio 是 GUI 优先（模型浏览器 + 参数面板 + 聊天界面），适合个人 \u002F 评估 \u002F 上手。Ollama 是 CLI \u002F Daemon 优先（后台跑 + REST API），适合应用嵌入 \u002F 脚本调用。两者都基于 llama.cpp，在 Mac M 系列上都已用 MLX。",{"q":503,"a":504},"支持 MLX 吗？","支持。Mac M1+ 上可加载 MLX 格式模型，速度比 GGUF + Metal 快 30–50%。模型搜索时筛选 MLX 即可。",{"q":506,"a":507},"OpenAI 兼容 API 怎么用？","开 Local Server → 默认端口 1234 → `http:\u002F\u002Flocalhost:1234\u002Fv1`。任何 OpenAI SDK 把 baseURL 改这个就能跑本地模型，零代码改动。",{"q":509,"a":510},"Headless 模式？","0.3+ 起支持 `lms server start` CLI 启动后台服务，无 GUI 即可跑 OpenAI 兼容 API，适合服务器 \u002F SSH 场景。",[512,513],"en","zh",{},true,"\u002Ftools\u002Fcoding\u002Flocal\u002Flm-studio","coding",[519,520,521],"windows","macos","linux",[523,527],{"plan":110,"price":524,"features":525,"notes":526},"免费","全功能 GUI + Headless API + GGUF\u002FMLX","供个人 \u002F 评估使用",{"plan":116,"price":528,"features":529,"notes":530},"联系咨询","团队部署 \u002F 商用 license","邮件 \u002F 官网联系","免费（个人 \u002F 评估） \u002F 企业 \u002F 商用咨询","2026-06-19",[534,535],"onboarding\u002Frag-pipeline-build","onboarding\u002Fclaude-code-getting-started",{"power":537,"ux":538,"price":538,"cn_support":480,"stability":537},4,5,{"title":12,"description":495},"coding\u002Flocal\u002Flm-studio",[542,545,547],{"name":543,"url":461,"accessed":544},"LM Studio 官网","2026-06-24",{"name":546,"url":469,"accessed":544},"Codersera — LM Studio Complete Guide 2026",{"name":548,"url":476,"accessed":544},"Codersera — Ollama vs LM Studio vs vLLM 2026","tools\u002Fcoding\u002Flocal\u002Flm-studio","本地 LLM 的 GUI 首选——模型浏览器 + GGUF\u002FMLX 推理 + OpenAI 兼容 API + Mac 原生优化",[493,552,553,554,555,556,557],"gui","gguf","mlx","llama-cpp","mac","openai-compatible","Mac \u002F Windows 桌面本地 LLM 的 GUI 首选——上手最快、模型浏览最舒服、自带 OpenAI 兼容 API。批量服务 \u002F 多用户场景用 vLLM；纯 CLI \u002F 嵌入应用走 Ollama。","https:\u002F\u002Flmstudio.ai","aNdpYFeU-2Jf8ElPrgRWVGVb70JZJ6MzrXXg5Tc6t3s",{"id":562,"title":563,"alternatives":564,"api_compatible":9,"body":565,"category":493,"chinese_friendly":538,"cover":1018,"description":1019,"domestic":496,"extension":497,"faq":1020,"free":496,"github":9,"languages":1033,"lastVerified":9,"meta":1034,"models":9,"navigation":515,"notSuitable":9,"opensource":515,"path":1035,"pillar":517,"platforms":1036,"priceTable":1039,"pricing":1047,"published":532,"relatedPlaybooks":1048,"relatedReviews":9,"score":1049,"self_host":515,"seo":1050,"seoTitle":9,"slug":17,"sources":1051,"stem":1058,"suitable":9,"tagline":1059,"tags":1060,"updated":544,"verdict":1067,"website":1068,"__hash__":1069},"tools\u002Ftools\u002Fcoding\u002Flocal\u002Flobe-chat.md","LobeChat",[16,15,14,540],{"type":19,"value":566,"toc":1006},[567,569,576,579,581,646,648,661,664,668,672,692,696,713,715,741,743,891,893,931,933,959,961,981,983],[22,568,25],{"id":24},[27,570,571,572,575],{},"LobeChat 是 LobeHub 团队的开源 AI 聊天框架，2023 年发布、GitHub 72k+ stars、MIT 协议。",[48,573,574],{},"Web + 桌面 + Docker 自托管三形态","，把 OpenAI \u002F Claude \u002F Gemini \u002F DeepSeek \u002F Qwen \u002F Kimi \u002F Ollama \u002F LM Studio 等 80+ 模型聚合到一个现代设计的客户端里。内置 RAG 知识库 + 插件市场 + 助手市场 + 多模型对比 + MCP，是当下综合最强的多模型 AI 客户端之一。",[27,577,578],{},"适合：需要 Web 端访问、Docker 自托管、多模型对比、丰富助手市场的用户；中文重度用户；想给团队 \u002F 家庭部署一个共享 AI 工作台。不适合：只用桌面 + 不需要 Web（Cherry Studio 同样优秀且更精细）、强企业 RBAC + 多租户（Open WebUI 多用户更完善）。",[22,580,40],{"id":40},[42,582,583,589,595,601,607,613,619,624,630,636],{},[45,584,585,588],{},[48,586,587],{},"多模型聚合","：OpenAI \u002F Claude \u002F Gemini \u002F DeepSeek \u002F Qwen \u002F Kimi \u002F 豆包 \u002F Groq \u002F Together \u002F OpenRouter \u002F Ollama \u002F LM Studio",[45,590,591,594],{},[48,592,593],{},"多模型对比","：同 prompt 给多模型并排回答",[45,596,597,600],{},[48,598,599],{},"本地 RAG 知识库","：上传 PDF \u002F Word \u002F 网页 → 向量化 → 检索引用",[45,602,603,606],{},[48,604,605],{},"插件市场","：联网搜索 \u002F 代码执行 \u002F 图像生成 \u002F 翻译等几十款官方插件",[45,608,609,612],{},[48,610,611],{},"助手市场","：几百个预设 AI 角色，一键导入",[45,614,615,618],{},[48,616,617],{},"MCP 协议","：扩展任意工具能力",[45,620,621],{},[48,622,623],{},"代码解释器 \u002F 文件上传 \u002F TTS \u002F 多模态",[45,625,626,629],{},[48,627,628],{},"Web + 桌面 + Docker","：三形态，数据可完全本地",[45,631,632,635],{},[48,633,634],{},"LobeHub Cloud","：官方云托管，免部署",[45,637,638,641,642,645],{},[48,639,640],{},"快捷指令 \u002F 工作流","：自定义 prompt 模板，",[31,643,644],{},"\u002Fpodcast-summary"," 类用法",[22,647,103],{"id":103},[42,649,650,656],{},[45,651,652,655],{},[48,653,654],{},"自托管 \u002F 桌面","：完全免费、MIT 开源",[45,657,658,660],{},[48,659,634],{},"：订阅制，云端托管 + 团队协作 + 同步",[27,662,663],{},"模型 API 费用按你自己的供应商付费；本地 Ollama \u002F LM Studio 零成本。",[22,665,667],{"id":666},"实测m2-自托管-docker连-openai-deepseek-本地-ollama","实测（M2 + 自托管 Docker，连 OpenAI + DeepSeek + 本地 Ollama）",[27,669,670],{},[48,671,132],{},[42,673,674,677,680,683,686,689],{},[45,675,676],{},"界面颜值是这一类工具里第一档（深色 \u002F 透明 \u002F 现代感）",[45,678,679],{},"多模型并排对比对选型极其有用：写一道复杂题，Claude \u002F GPT \u002F DeepSeek 直接对比答案",[45,681,682],{},"知识库 RAG 上传 50+ PDF 后检索准确，引用片段可视化",[45,684,685],{},"助手市场拿来即用——「Code Reviewer」「Translation Polish」节省 prompt 编写",[45,687,688],{},"Docker 一键部署，团队 5 人共享流畅",[45,690,691],{},"多平台数据同步（Cloud \u002F WebDAV）",[27,693,694],{},[48,695,154],{},[42,697,698,701,704,707,710],{},[45,699,700],{},"自托管要熟悉 Docker + 反代 + HTTPS",[45,702,703],{},"国内连 OpenAI \u002F Claude 需自带网络方案",[45,705,706],{},"Web 版数据存 LobeHub，隐私敏感场景走桌面 \u002F Docker",[45,708,709],{},"插件市场质量参差，要自筛",[45,711,712],{},"团队多人共享需配 LobeHub Cloud 或自建数据库（Postgres + S3）",[22,714,174],{"id":174},[176,716,717,720,726,729,732,735,738],{},[45,718,719],{},"选形态：Web（chat.lobehub.com 注册即用） \u002F 桌面（GitHub Releases 下载） \u002F Docker",[45,721,722,723],{},"Docker：",[31,724,725],{},"docker run -d -p 3210:3210 -e OPENAI_API_KEY=sk-xxx --name lobe-chat lobehub\u002Flobe-chat",[45,727,728],{},"设置 → AI 服务商 → 添加 OpenAI \u002F Claude \u002F DeepSeek \u002F Ollama",[45,730,731],{},"模型选择器测试对话",[45,733,734],{},"知识库：拖文件 → 等向量化 → 对话引用",[45,736,737],{},"助手市场拉「Code Reviewer」「论文翻译润色」试用",[45,739,740],{},"进阶：插件市场启用联网搜索 \u002F 代码执行；MCP 自定义工具",[22,742,204],{"id":204},[206,744,745,760],{},[209,746,747],{},[212,748,749,751,753,756,758],{},[215,750,217],{},[215,752,563],{},[215,754,755],{},"Cherry Studio",[215,757,225],{},[215,759,12],{},[230,761,762,776,790,803,817,833,847,861,877],{},[212,763,764,766,768,771,774],{},[235,765,237],{},[235,767,628],{},[235,769,770],{},"桌面",[235,772,773],{},"Docker \u002F 桌面",[235,775,770],{},[212,777,778,780,783,785,787],{},[235,779,587],{},[235,781,782],{},"✅ 80+",[235,784,274],{},[235,786,274],{},[235,788,789],{},"本地为主",[212,791,792,794,797,799,801],{},[235,793,593],{},[235,795,796],{},"✅ 一等",[235,798,274],{},[235,800,320],{},[235,802,320],{},[212,804,805,808,810,812,815],{},[235,806,807],{},"知识库 RAG",[235,809,274],{},[235,811,274],{},[235,813,814],{},"✅ + oikb",[235,816,320],{},[212,818,819,822,825,828,831],{},[235,820,821],{},"插件 \u002F 助手市场",[235,823,824],{},"✅ 丰富",[235,826,827],{},"300+ 助手",[235,829,830],{},"Tools",[235,832,320],{},[212,834,835,838,840,842,845],{},[235,836,837],{},"MCP",[235,839,274],{},[235,841,274],{},[235,843,844],{},"✅ mcpo",[235,846,320],{},[212,848,849,852,855,857,859],{},[235,850,851],{},"多用户",[235,853,854],{},"配 Cloud \u002F 自建",[235,856,263],{},[235,858,796],{},[235,860,263],{},[212,862,863,866,869,872,875],{},[235,864,865],{},"GitHub Stars",[235,867,868],{},"72k+",[235,870,871],{},"60k+",[235,873,874],{},"126k+",[235,876,310],{},[212,878,879,882,884,887,889],{},[235,880,881],{},"开源协议",[235,883,338],{},[235,885,886],{},"AGPL-3.0",[235,888,338],{},[235,890,335],{},[22,892,361],{"id":361},[42,894,895,901,907,913,919,925],{},[45,896,897,900],{},[48,898,899],{},"Web 版数据不本地","：隐私敏感选桌面或 Docker 自托管",[45,902,903,906],{},[48,904,905],{},"国内连海外模型走中转","：直连 OpenAI \u002F Claude 不稳，配 OpenRouter \u002F Ofox \u002F 国内中转",[45,908,909,912],{},[48,910,911],{},"Docker 自托管暴露公网","：上反代 + HTTPS + Auth + 备份数据库",[45,914,915,918],{},[48,916,917],{},"嵌入模型中文优化","：默认嵌入对中文一般，配 bge-m3 \u002F 硅基流动 Pro 版",[45,920,921,924],{},[48,922,923],{},"插件市场审一遍","：第三方插件可执行代码，团队部署谨慎启用",[45,926,927,930],{},[48,928,929],{},"同步选 Cloud vs WebDAV","：团队多端走 LobeHub Cloud；个人多设备 WebDAV 即可",[22,932,397],{"id":396},[42,934,935,938,941,944,947,950,953,956],{},[45,936,937],{},"✅ Web + 桌面双形态需求",[45,939,940],{},"✅ Docker 自托管 \u002F 团队共享",[45,942,943],{},"✅ 多模型对比 \u002F 选型",[45,945,946],{},"✅ 中文重度用户",[45,948,949],{},"✅ 助手市场 \u002F 插件生态用户",[45,951,952],{},"❌ 强企业 RBAC + 多租户（Open WebUI 更完善）",[45,954,955],{},"❌ 只要桌面 + 数据完全本地（Cherry Studio 同样优秀）",[45,957,958],{},"❌ 完全不会碰 Docker",[22,960,423],{"id":423},[42,962,963,967,971,975],{},[45,964,965],{},[429,966,444],{"href":443},[45,968,969],{},[429,970,438],{"href":437},[45,972,973],{},[429,974,432],{"href":431},[45,976,977],{},[429,978,980],{"href":979},"\u002Fplaybook\u002Fonboarding\u002Frag-pipeline-build","RAG Pipeline 搭建 Playbook",[22,982,453],{"id":453},[176,984,985,992,999],{},[45,986,987,988],{},"LobeChat GitHub 仓库（72k+ stars，MIT）",[429,989,990],{"href":990,"rel":991},"https:\u002F\u002Fgithub.com\u002Flobehub\u002Flobe-chat",[463],[45,993,994,995],{},"腾讯云开发者社区 — Lobe Chat 本地化 AI 聊天终极桌面客户端（2026-01）",[429,996,997],{"href":997,"rel":998},"https:\u002F\u002Fcloud.tencent.com\u002Fdeveloper\u002Farticle\u002F2622150",[463],[45,1000,1001,1002],{},"Ofox.ai — LobeChat 完全配置指南 2026（2026-04-17）",[429,1003,1004],{"href":1004,"rel":1005},"https:\u002F\u002Fofox.ai\u002Fzh\u002Fblog\u002Flobechat-api-configuration-guide-2026",[463],{"title":479,"searchDepth":480,"depth":480,"links":1007},[1008,1009,1010,1011,1012,1013,1014,1015,1016,1017],{"id":24,"depth":483,"text":25},{"id":40,"depth":483,"text":40},{"id":103,"depth":483,"text":103},{"id":666,"depth":483,"text":667},{"id":174,"depth":483,"text":174},{"id":204,"depth":483,"text":204},{"id":361,"depth":483,"text":361},{"id":396,"depth":483,"text":397},{"id":423,"depth":483,"text":423},{"id":453,"depth":483,"text":453},"\u002Fimg\u002Ftools\u002Flobe-chat.webp","LobeChat 真实评测：LobeHub 团队开源 AI 聊天框架，GitHub 72k+ stars、MIT 协议。Web + 桌面（Win\u002FMac\u002FLinux\u002FDocker）双形态，支持 OpenAI \u002F Claude \u002F Gemini \u002F DeepSeek \u002F Qwen \u002F Ollama 等 80+ 模型，内置 RAG 知识库 + 插件市场 + 助手市场 + 多模型对比。",[1021,1024,1027,1030],{"q":1022,"a":1023},"Web 版 vs 桌面版 vs Docker 自托管，怎么选？","Web 版（chat.lobehub.com）最快上手但数据存 LobeHub 服务器；桌面版数据本地存、隐私好；Docker 自托管对团队 \u002F 公司部署最优，完全掌控数据。",{"q":1025,"a":1026},"支持哪些模型？","80+ 模型：OpenAI 全系列、Anthropic Claude、Google Gemini、DeepSeek、Qwen、Kimi、Moonshot、字节豆包、Groq、Together、OpenRouter、Ollama \u002F LM Studio 本地模型，以及任何 OpenAI 兼容 API。",{"q":1028,"a":1029},"多模型对比怎么用？","同一对话窗口里把消息广播给多个模型并排回答，选型 \u002F 评估特别有用——直接看 Claude 和 GPT 在同一 prompt 下的回答差异。",{"q":1031,"a":1032},"助手市场是什么？","LobeHub 维护的预设 AI 角色市场（代码审查 \u002F 翻译 \u002F 写作 \u002F 角色扮演等几百个），一键拉到本地用，省去自己写 System Prompt。",[513,512],{},"\u002Ftools\u002Fcoding\u002Flocal\u002Flobe-chat",[1037,519,520,521,1038],"web","docker",[1040,1043],{"plan":654,"price":524,"features":1041,"notes":1042},"全功能 \u002F 80+ 模型 \u002F 知识库 \u002F 插件 \u002F 助手市场","MIT 协议",{"plan":634,"price":1044,"features":1045,"notes":1046},"订阅制","云端托管 \u002F 免部署 \u002F 团队协作 \u002F 同步","chat.lobehub.com 注册即用","完全免费（MIT 开源） \u002F LobeHub Cloud 订阅",[534,535],{"power":538,"ux":538,"price":538,"cn_support":538,"stability":537},{"title":563,"description":1019},[1052,1054,1056],{"name":1053,"url":990,"accessed":544},"LobeChat GitHub",{"name":1055,"url":997,"accessed":544},"腾讯云开发者社区 — Lobe Chat 终极桌面客户端",{"name":1057,"url":1004,"accessed":544},"Ofox.ai — LobeChat 完全配置指南 2026","tools\u002Fcoding\u002Flocal\u002Flobe-chat","现代设计的开源 AI 聊天框架——Web + 桌面双形态、72k+ stars、多模型 + 知识库 + 插件市场",[493,1037,1061,1062,1063,1064,1065,1066],"desktop","multi-model","rag","plugin","mcp","open-source","颜值与功能双优的多模型 AI 聊天客户端。要 Web + 桌面双形态、自托管 Docker、多模型对比、丰富助手市场——LobeChat 是综合最强；纯桌面体验 Cherry Studio 同样优秀。","https:\u002F\u002Flobehub.com","pm20AeqHCiMi5MF0JALNWKY72MAdrs-uWEpbDmfEKbY",1784565442423]