[{"data":1,"prerenderedAt":1344},["ShallowReactive",2],{"header-counts":3,"footer-counts":6,"compare-open-interpreter-vs-openclaw":8,"compare-a-open-interpreter":9,"compare-b-openclaw":564},{"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":527,"chinese_friendly":516,"cover":528,"description":529,"domestic":530,"extension":531,"faq":8,"free":530,"github":508,"languages":532,"lastVerified":534,"meta":535,"models":8,"navigation":536,"notSuitable":8,"opensource":536,"path":537,"pillar":538,"platforms":539,"priceTable":8,"pricing":543,"published":544,"relatedPlaybooks":8,"relatedReviews":8,"score":545,"self_host":530,"seo":548,"seoTitle":549,"slug":550,"sources":551,"stem":554,"suitable":8,"tagline":555,"tags":556,"updated":534,"verdict":562,"website":500,"__hash__":563},"tools\u002Ftools\u002Fagent\u002Fdesktop\u002Fopen-interpreter.md","Open Interpreter",[13,14,15],"agent\u002Fdesktop\u002Fclaude-desktop","agent\u002Fgeneral\u002Fopenmanus","agent\u002Fdesktop\u002Fopenclaw",{"type":17,"value":18,"toc":511},"minimark",[19,24,33,36,39,96,99,102,106,116,121,144,149,169,172,210,213,348,351,402,406,435,439,445,451,461,467,470,486,489,494],[20,21,23],"h2",{"id":22},"tldr","TL;DR",[25,26,27,28,32],"p",{},"Open Interpreter 是一个开源的本地代码执行 Agent（AGPL 协议），让 LLM 直接在你的终端里写代码、跑代码、读写文件、执行系统命令、操控浏览器。它不像 ChatGPT 那样在沙箱里跑代码——它在你真实的机器上执行，有完整的系统权限。支持 GPT-4 \u002F Claude \u002F 本地模型（Ollama \u002F LM Studio），Python 实现，",[29,30,31],"code",{},"pip install"," 即用。",[25,34,35],{},"适合：技术用户做本地自动化（批量文件处理、数据分析、系统管理）、探索性编程（让 AI 跑代码看结果再迭代）、原型开发。不适合：生产环境（稳定性不够）、非技术用户（安全风险高）、敏感数据环境（LLM 能读你所有文件）、商业闭源项目（AGPL 传染性协议）。",[20,37,38],{"id":38},"核心能力",[40,41,42,50,56,62,68,74,80,90],"ul",{},[43,44,45,49],"li",{},[46,47,48],"strong",{},"本地代码执行","：LLM 生成 Python \u002F JS \u002F Shell 代码，直接在你的机器上跑，看到真实输出",[43,51,52,55],{},[46,53,54],{},"文件系统操作","：读写文件、创建目录、搜索文件内容、批量重命名，完整的文件系统访问",[43,57,58,61],{},[46,59,60],{},"系统命令执行","：跑任意 shell 命令——安装软件、配置环境、管理进程、操作系统",[43,63,64,67],{},[46,65,66],{},"浏览器控制","：通过 Playwright \u002F Selenium 操控浏览器，自动化网页操作、数据抓取",[43,69,70,73],{},[46,71,72],{},"多模型支持","：OpenAI \u002F Anthropic \u002F Google \u002F Ollama \u002F LM Studio \u002F 任何 OpenAI 兼容端点",[43,75,76,79],{},[46,77,78],{},"交互式 REPL","：终端对话式交互，自然语言描述任务，Agent 自主分解 + 执行 + 反馈",[43,81,82,85,86,89],{},[46,83,84],{},"执行确认机制","：每步代码执行前可选确认（",[29,87,88],{},"--auto_run"," 跳过），防止误操作",[43,91,92,95],{},[46,93,94],{},"多语言执行","：不限于 Python——JS、Shell、R、Ruby 等按需切换",[20,97,98],{"id":98},"价格",[25,100,101],{},"完全免费，AGPL-3.0 开源。成本在于 LLM API 调用费（用 GPT-4 \u002F Claude）或本地模型硬件成本（用 Ollama 免费）。",[20,103,105],{"id":104},"体验与评测资料整理","体验与评测（资料整理）",[107,108,109],"blockquote",{},[25,110,111,112,115],{},"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",[46,113,114],{},"典型场景","：自动化文件处理、数据分析、本地脚本执行等任务（配云 API 或本地 Ollama 均可，硬件门槛低）。",[25,117,118],{},[46,119,120],{},"亮点：",[40,122,123,129,132,135,138,141],{},[43,124,125,128],{},[29,126,127],{},"interpreter"," 一行启动，配 API key 即用，对话式交互体验流畅",[43,130,131],{},"让它「把这个文件夹所有 CSV 合并去重，按日期排序，生成图表」——写 Python → 跑 → 看报错 → 改 → 再跑，全自主完成",[43,133,134],{},"文件操作强——搜索、批量重命名、格式转换，比手写脚本快",[43,136,137],{},"数据分析场景好用——让它读 Excel、清洗数据、画 matplotlib 图，迭代式探索很方便",[43,139,140],{},"配 Ollama 本地模型完全离线运行，隐私敏感任务不碰云",[43,142,143],{},"浏览器自动化——让它打开网页、填表、截图，比手动写 Playwright 脚本快",[25,145,146],{},[46,147,148],{},"踩坑：",[40,150,151,154,157,160,163,166],{},[43,152,153],{},"安全风险是最大问题——LLM 能跑任意命令，一次让它「清理临时文件」差点删了项目目录，幸好有确认机制拦住",[43,155,156],{},"稳定性一般——复杂任务偶尔死循环（反复跑失败代码）、MCP 连接断开、进程卡死",[43,158,159],{},"长任务上下文丢失——对话太长后 Agent 忘记前面的指令，需要手动提醒",[43,161,162],{},"本地模型效果有限——Qwen3 7B 能做简单任务，复杂逻辑推理要上 Claude \u002F GPT-4",[43,164,165],{},"AGPL 协议限制商业使用——如果你在公司用且分发，要注意开源传染性问题",[43,167,168],{},"Windows 路径问题——部分 Shell 命令在 Windows 有兼容性 bug，Mac \u002F Linux 更稳",[20,170,171],{"id":171},"上手",[173,174,175,181,192,198,201,207],"ol",{},[43,176,177,178],{},"安装：",[29,179,180],{},"pip install open-interpreter",[43,182,183,184,187,188,191],{},"配模型：",[29,185,186],{},"interpreter --model claude-3-5-sonnet","（或 ",[29,189,190],{},"--model ollama\u002Fqwen3-coder:7b"," 用本地）",[43,193,194,195,197],{},"启动：",[29,196,127],{}," 进入交互式 REPL",[43,199,200],{},"描述任务：「把 ~\u002FDownloads 所有图片转成 webp，压缩到 80% 质量」",[43,202,203,204,206],{},"确认执行——每步代码执行前会提示确认（开发阶段别开 ",[29,205,88],{},"）",[43,208,209],{},"进阶：配自定义工具 \u002F MCP server 扩展 Agent 能力",[20,211,212],{"id":212},"对比",[214,215,216,237],"table",{},[217,218,219],"thead",{},[220,221,222,226,228,231,234],"tr",{},[223,224,225],"th",{},"维度",[223,227,11],{},[223,229,230],{},"Claude Desktop",[223,232,233],{},"OpenManus",[223,235,236],{},"Devika",[238,239,240,257,274,289,303,317,333],"tbody",{},[220,241,242,246,249,252,255],{},[243,244,245],"td",{},"形态",[243,247,248],{},"CLI",[243,250,251],{},"Desktop App",[243,253,254],{},"Web",[243,256,254],{},[220,258,259,262,265,268,271],{},[243,260,261],{},"开源",[243,263,264],{},"✅ AGPL",[243,266,267],{},"❌",[243,269,270],{},"✅ Apache 2.0",[243,272,273],{},"✅ MIT",[220,275,276,279,282,285,287],{},[243,277,278],{},"代码执行",[243,280,281],{},"✅ 本地真实",[243,283,284],{},"✅ 沙箱",[243,286,284],{},[243,288,284],{},[220,290,291,293,296,299,301],{},[243,292,66],{},[243,294,295],{},"✅",[243,297,298],{},"✅ Computer Use",[243,300,295],{},[243,302,295],{},[220,304,305,308,311,313,315],{},[243,306,307],{},"本地模型",[243,309,310],{},"✅ Ollama",[243,312,267],{},[243,314,295],{},[243,316,295],{},[220,318,319,322,325,328,331],{},[243,320,321],{},"安全性",[243,323,324],{},"⚠️ 低（真实执行）",[243,326,327],{},"✅ 高（沙箱）",[243,329,330],{},"✅ 高",[243,332,330],{},[220,334,335,338,341,344,346],{},[243,336,337],{},"稳定性",[243,339,340],{},"中",[243,342,343],{},"高",[243,345,340],{},[243,347,340],{},[20,349,350],{"id":350},"避坑",[40,352,353,362,372,378,384,390,396],{},[43,354,355,358,359,361],{},[46,356,357],{},"一定要开确认机制","：默认每步执行前确认，别图省事开 ",[29,360,88],{},"，LLM 误删文件的代价远大于点几下确认",[43,363,364,367,368,371],{},[46,365,366],{},"别在生产服务器跑","：Open Interpreter 有完整系统权限，LLM 幻觉可能跑出 ",[29,369,370],{},"rm -rf"," \u002F 覆盖配置文件等灾难操作",[43,373,374,377],{},[46,375,376],{},"敏感数据隔离","：LLM 能读你所有文件，处理含密钥 \u002F 密码 \u002F 用户数据的环境要慎重，建议在 Docker 容器里跑",[43,379,380,383],{},[46,381,382],{},"长任务分段","：对话超过 20-30 轮后上下文退化，复杂任务拆成小段，每段确认结果再继续",[43,385,386,389],{},[46,387,388],{},"本地模型别期望太高","：7B\u002F14B 模型能做文件操作和简单脚本，复杂逻辑推理和调试要上 Claude \u002F GPT-4",[43,391,392,395],{},[46,393,394],{},"AGPL 协议注意","：商业项目分发时 AGPL 有传染性，内部使用没问题，分发 \u002F SaaS 化要咨询法务",[43,397,398,401],{},[46,399,400],{},"Windows 用 WSL","：原生 Windows 有路径 \u002F 权限 \u002F Shell 兼容问题，WSL 下体验一致",[20,403,405],{"id":404},"适合-不适合","适合 \u002F 不适合",[40,407,408,411,414,417,420,423,426,429,432],{},[43,409,410],{},"✅ 技术用户做本地自动化（文件批处理、数据分析、系统管理）",[43,412,413],{},"✅ 探索性编程（让 AI 跑代码看结果再迭代）",[43,415,416],{},"✅ 原型开发 \u002F 快速脚本生成",[43,418,419],{},"✅ 想用本地模型（Ollama）做隐私敏感的自动化",[43,421,422],{},"✅ 学习 \u002F 研究本地 Agent 架构",[43,424,425],{},"❌ 生产环境（稳定性不够 + 安全风险高）",[43,427,428],{},"❌ 非技术用户（误操作风险大，确认机制需要判断力）",[43,430,431],{},"❌ 敏感数据环境（LLM 能访问所有文件）",[43,433,434],{},"❌ 商业闭源项目分发（AGPL 传染性协议）",[20,436,438],{"id":437},"faq","FAQ",[25,440,441,444],{},[46,442,443],{},"Q: Open Interpreter 安全吗？","\nA: 本身不安全——它在你真实机器上执行任意代码，LLM 幻觉可能导致误删文件 \u002F 执行危险命令。缓解措施：开确认机制、在 Docker 容器跑、别给 root 权限、敏感数据隔离。生产环境不推荐。",[25,446,447,450],{},[46,448,449],{},"Q: 和 Claude Desktop 的 Computer Use 区别？","\nA: Claude Desktop 在沙箱虚拟机里操作，安全隔离好。Open Interpreter 在你真实机器上执行，能力更强但风险更高。追求安全选 Claude Desktop，追求能力和免费选 Open Interpreter。",[25,452,453,456,457,460],{},[46,454,455],{},"Q: 能用本地模型吗？","\nA: 可以。",[29,458,459],{},"interpreter --model ollama\u002Fqwen3-coder:7b"," 即用 Ollama 本地模型，完全不碰云 API。简单任务（文件操作、基础脚本）够用，复杂推理和调试建议上 Claude \u002F GPT-4。",[25,462,463,466],{},[46,464,465],{},"Q: 和 AutoGPT \u002F BabyAGI 区别？","\nA: AutoGPT \u002F BabyAGI 偏「自主规划 + 工具调用」的通用 Agent，工具链固定。Open Interpreter 偏「代码执行」——它通过写代码来完成任务，灵活性更高（任何能写代码实现的都能做），但安全风险也更高。",[20,468,469],{"id":469},"相关阅读",[25,471,472,477,478,477,482],{},[473,474,476],"a",{"href":475},"\u002Fcoding\u002Flocal\u002Fjan.html","Jan"," · ",[473,479,481],{"href":480},"\u002Fcoding\u002Flocal\u002Fgpt4all.html","GPT4All",[473,483,485],{"href":484},"\u002Fcoding\u002Flocal\u002Fvllm.html","vLLM",[20,487,488],{"id":488},"来源",[107,490,491],{},[25,492,493],{},"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。",[40,495,496,504],{},[43,497,498],{},[473,499,503],{"href":500,"rel":501},"https:\u002F\u002Fopeninterpreter.com",[502],"nofollow","官网",[43,505,506],{},[473,507,510],{"href":508,"rel":509},"https:\u002F\u002Fgithub.com\u002FOpenInterpreter\u002Fopen-interpreter",[502],"GitHub",{"title":512,"searchDepth":513,"depth":513,"links":514},"",3,[515,517,518,519,520,521,522,523,524,525,526],{"id":22,"depth":516,"text":23},2,{"id":38,"depth":516,"text":38},{"id":98,"depth":516,"text":98},{"id":104,"depth":516,"text":105},{"id":171,"depth":516,"text":171},{"id":212,"depth":516,"text":212},{"id":350,"depth":516,"text":350},{"id":404,"depth":516,"text":405},{"id":437,"depth":516,"text":438},{"id":469,"depth":516,"text":469},{"id":488,"depth":516,"text":488},"desktop","\u002Fimg\u002Ftools\u002Fopen-interpreter.webp","Open Interpreter 真实评测：开源本地代码执行 Agent（AGPL 协议），让 LLM 在终端直接跑 Python\u002FJS\u002FShell 代码，操控浏览器、读写文件、执行系统命令。支持 GPT-4\u002FClaude\u002F本地模型，适合需要 LLM 实时操作本地环境的开发者。",false,"md",[533],"en","2026-07-30",{},true,"\u002Ftools\u002Fagent\u002Fdesktop\u002Fopen-interpreter","agent",[540,541,542],"windows","macos","linux","Free \u002F 开源（AGPL）","2026-07-05",{"power":546,"ux":513,"price":547,"cn_support":516,"stability":513},4,5,{"title":11,"description":529},"Open Interpreter - 本地代码执行 Agent 评测 | AIHO","agent\u002Fdesktop\u002Fopen-interpreter",[552,553],{"title":503,"url":500},{"title":510,"url":508},"tools\u002Fagent\u002Fdesktop\u002Fopen-interpreter","本地代码执行 Agent，让 LLM 在你的机器上跑代码",[557,558,559,560,561],"desktop-agent","opensource","code-execution","local","python","开创性的本地代码执行 Agent，让 LLM 直接操控你的电脑跑代码、读写文件、操作浏览器，适合技术用户做自动化和探索；安全风险高、稳定性一般，不适合生产环境或非技术用户。","L4JM70girxbUKgBD0dOIoVvQD-2eCVHnpHH7ZtdCUYo",{"id":565,"title":566,"alternatives":567,"api_compatible":8,"body":570,"category":527,"chinese_friendly":546,"cover":1289,"description":1290,"domestic":530,"extension":531,"faq":1291,"free":530,"github":8,"languages":1304,"lastVerified":8,"meta":1307,"models":8,"navigation":536,"notSuitable":8,"opensource":536,"path":1308,"pillar":538,"platforms":1309,"priceTable":1311,"pricing":1315,"published":1316,"relatedPlaybooks":1317,"relatedReviews":8,"score":1319,"self_host":536,"seo":1320,"seoTitle":1321,"slug":15,"sources":1322,"stem":1332,"suitable":8,"tagline":1333,"tags":1334,"updated":1325,"verdict":1341,"website":1342,"__hash__":1343},"tools\u002Ftools\u002Fagent\u002Fdesktop\u002Fopenclaw.md","OpenClaw",[13,568,569],"agent\u002Fdesktop\u002Fautoglm","agent\u002Fdesktop\u002Fcrush",{"type":17,"value":571,"toc":1277},[572,574,577,580,582,650,652,666,670,674,697,701,727,729,916,919,951,953,1127,1129,1191,1193,1219,1221,1241,1243,1273],[20,573,23],{"id":22},[25,575,576],{},"OpenClaw 是 2026 现象级开源个人 AI Agent 框架（MIT 协议）——Peter Steinberger 2025 末创建，2026-02 移交开源基金会，60 天 28-35.5 万 GitHub stars 历史级速度。差异点：Gateway 网关连接 20+ IM 平台（Discord \u002F Slack \u002F Telegram \u002F WhatsApp \u002F iMessage \u002F Matrix \u002F Signal \u002F Microsoft Teams \u002F Zalo \u002F 微信 \u002F 钉钉 \u002F 飞书）+ 30 分钟 Heartbeat 主动触发 + Cron 调度 + ClawHub 5400+ Skills 市场 + 多 Agent 隔离 + 本地长期记忆。完全自托管，本地 Ollama 模型零成本运行。",[25,578,579],{},"适合：要『手机 IM 一句话支使 AI 干活』的 power user；监控告警 \u002F 定时简报 \u002F 邮件分拣 \u002F 浏览器自动化等 7×24 后台场景；数据隐私敏感 + 自托管刚需；中国大陆用户（本地 Ollama 全免费）。不适合：纯 GUI \u002F 桌面交互党（Claude Desktop 更顺）；不会折腾 Node \u002F Docker \u002F WSL 的纯用户；要『一句话就跑通』的非技术用户。",[20,581,38],{"id":38},[40,583,584,590,596,602,608,614,620,626,632,638,644],{},[43,585,586,589],{},[46,587,588],{},"Gateway 多渠道","：Discord \u002F Slack \u002F Telegram \u002F WhatsApp \u002F iMessage \u002F Matrix \u002F Signal \u002F MS Teams \u002F Zalo \u002F 微信 \u002F 钉钉 \u002F 飞书 + WebChat + 移动 node",[43,591,592,595],{},[46,593,594],{},"Skills 插件系统","：YAML 描述 + 一键安装 + ClawHub 5400+ 社区市场",[43,597,598,601],{},[46,599,600],{},"多 Agent 隔离","：work \u002F personal \u002F test 角色独立记忆 + 任务",[43,603,604,607],{},[46,605,606],{},"30 分钟 Heartbeat","：真正 7×24 主动运行",[43,609,610,613],{},[46,611,612],{},"Cron 调度","：自然语言生成定时任务，无需写 Cron 表达式",[43,615,616,619],{},[46,617,618],{},"本地长期记忆","：~\u002F.openclaw\u002Fmemory\u002Flong-term.json",[43,621,622,625],{},[46,623,624],{},"沙箱权限模式","：区分敏感操作权限，避免误操作",[43,627,628,631],{},[46,629,630],{},"多 LLM provider","：GPT \u002F Claude \u002F Gemini \u002F 本地 Ollama \u002F Z.AI \u002F OpenRouter",[43,633,634,637],{},[46,635,636],{},"MCP 兼容","：可调用 MCP server",[43,639,640,643],{},[46,641,642],{},"Web Control UI","：浏览器仪表盘（chat \u002F config \u002F sessions \u002F nodes）",[43,645,646,649],{},[46,647,648],{},"macOS app + iOS \u002F Android 移动 node","：跨设备触发",[20,651,98],{"id":98},[40,653,654,660,663],{},[43,655,656,659],{},[46,657,658],{},"Self-host OSS","：$0；MIT 完全开放",[43,661,662],{},"真实成本 = LLM API（GPT \u002F Claude 走 API 计费）或 $0（本地 Ollama）",[43,664,665],{},"轻度 $5-10 \u002F 月，中度 $20-30，重度 $50+ —— 全部来自外部 LLM 而非 OpenClaw 本身",[20,667,669],{"id":668},"实测个人-724-监控告警-简报","实测（个人 7×24 + 监控告警 + 简报）",[25,671,672],{},[46,673,120],{},[40,675,676,679,682,685,688,691,694],{},[43,677,678],{},"IM 一句话支使 AI 体验是真『未来感』，离开电脑也能干活",[43,680,681],{},"Heartbeat 让定时任务 + 监控钩子真正后台运行，不靠操作系统 Cron",[43,683,684],{},"ClawHub 5400+ Skills 覆盖几乎所有日常需求",[43,686,687],{},"多 Agent 隔离让工作 \u002F 生活 prompt 不串",[43,689,690],{},"本地 Ollama 路径让数据 + 隐私完全自主",[43,692,693],{},"微信 \u002F 钉钉 \u002F 飞书 plugin 让国内用户无网络 \u002F 支付门槛",[43,695,696],{},"MIT 协议 + 社区驱动，企业内部部署合规友好",[25,698,699],{},[46,700,148],{},[40,702,703,706,709,712,715,718,721,724],{},[43,704,705],{},"Node 24 推荐 \u002F 22 LTS 兜底，Node 18 已经勉强（最新 Skill 偶有兼容问题）",[43,707,708],{},"Skill 质量参差，社区 5400+ 里要挑信任作者",[43,710,711],{},"国内 IM 渠道（微信 \u002F 钉钉 \u002F 飞书）的 plugin 受平台风控影响，账号风险要评估",[43,713,714],{},"多 Agent 隔离的 memory 文件可能膨胀，要定期清理",[43,716,717],{},"Heartbeat 频繁触发 LLM 调用会推高 API 费用，预算控制要配 budget cap",[43,719,720],{},"Web Control UI 在大量 session 下渲染偏慢",[43,722,723],{},"沙箱模式默认权限保守，复杂自动化要手动放权",[43,725,726],{},"Windows 纯生兼容差，建议 WSL2 \u002F Docker 路径",[20,728,171],{"id":171},[730,731,735],"pre",{"className":732,"code":733,"language":734,"meta":512,"style":512},"language-bash shiki shiki-themes github-light github-dark","# 方法一：NPM\nnpm install -g openclaw\nopenclaw --version\n\n# 方法二：Docker\ndocker run -d --name openclaw \\\n  -v ~\u002F.openclaw:\u002Froot\u002F.openclaw \\\n  openclaw\u002Fopenclaw:latest\n\n# 方法三：源码\ngit clone https:\u002F\u002Fgithub.com\u002Fopenclaw\u002Fopenclaw.git\ncd openclaw && npm install && npm run start\n\n# 引导配置（model \u002F channel \u002F 权限）\nopenclaw onboard\n\n# 启动\nopenclaw                 # CLI\nopenclaw dashboard       # Web UI\n","bash",[29,736,737,746,763,771,776,781,802,813,819,824,830,842,867,872,878,886,891,897,905],{"__ignoreMap":512},[738,739,742],"span",{"class":740,"line":741},"line",1,[738,743,745],{"class":744},"sJ8bj","# 方法一：NPM\n",[738,747,748,752,756,760],{"class":740,"line":516},[738,749,751],{"class":750},"sScJk","npm",[738,753,755],{"class":754},"sZZnC"," install",[738,757,759],{"class":758},"sj4cs"," -g",[738,761,762],{"class":754}," openclaw\n",[738,764,765,768],{"class":740,"line":513},[738,766,767],{"class":750},"openclaw",[738,769,770],{"class":758}," --version\n",[738,772,773],{"class":740,"line":546},[738,774,775],{"emptyLinePlaceholder":536},"\n",[738,777,778],{"class":740,"line":547},[738,779,780],{"class":744},"# 方法二：Docker\n",[738,782,784,787,790,793,796,799],{"class":740,"line":783},6,[738,785,786],{"class":750},"docker",[738,788,789],{"class":754}," run",[738,791,792],{"class":758}," -d",[738,794,795],{"class":758}," --name",[738,797,798],{"class":754}," openclaw",[738,800,801],{"class":758}," \\\n",[738,803,805,808,811],{"class":740,"line":804},7,[738,806,807],{"class":758},"  -v",[738,809,810],{"class":754}," ~\u002F.openclaw:\u002Froot\u002F.openclaw",[738,812,801],{"class":758},[738,814,816],{"class":740,"line":815},8,[738,817,818],{"class":754},"  openclaw\u002Fopenclaw:latest\n",[738,820,822],{"class":740,"line":821},9,[738,823,775],{"emptyLinePlaceholder":536},[738,825,827],{"class":740,"line":826},10,[738,828,829],{"class":744},"# 方法三：源码\n",[738,831,833,836,839],{"class":740,"line":832},11,[738,834,835],{"class":750},"git",[738,837,838],{"class":754}," clone",[738,840,841],{"class":754}," https:\u002F\u002Fgithub.com\u002Fopenclaw\u002Fopenclaw.git\n",[738,843,845,848,850,854,856,858,860,862,864],{"class":740,"line":844},12,[738,846,847],{"class":758},"cd",[738,849,798],{"class":754},[738,851,853],{"class":852},"sVt8B"," && ",[738,855,751],{"class":750},[738,857,755],{"class":754},[738,859,853],{"class":852},[738,861,751],{"class":750},[738,863,789],{"class":754},[738,865,866],{"class":754}," start\n",[738,868,870],{"class":740,"line":869},13,[738,871,775],{"emptyLinePlaceholder":536},[738,873,875],{"class":740,"line":874},14,[738,876,877],{"class":744},"# 引导配置（model \u002F channel \u002F 权限）\n",[738,879,881,883],{"class":740,"line":880},15,[738,882,767],{"class":750},[738,884,885],{"class":754}," onboard\n",[738,887,889],{"class":740,"line":888},16,[738,890,775],{"emptyLinePlaceholder":536},[738,892,894],{"class":740,"line":893},17,[738,895,896],{"class":744},"# 启动\n",[738,898,900,902],{"class":740,"line":899},18,[738,901,767],{"class":750},[738,903,904],{"class":744},"                 # CLI\n",[738,906,908,910,913],{"class":740,"line":907},19,[738,909,767],{"class":750},[738,911,912],{"class":754}," dashboard",[738,914,915],{"class":744},"       # Web UI\n",[25,917,918],{},"配本地免费 Ollama：",[730,920,922],{"className":732,"code":921,"language":734,"meta":512,"style":512},"ollama pull llama3\nopenclaw config set ai.provider \"ollama\"\n",[29,923,924,935],{"__ignoreMap":512},[738,925,926,929,932],{"class":740,"line":741},[738,927,928],{"class":750},"ollama",[738,930,931],{"class":754}," pull",[738,933,934],{"class":754}," llama3\n",[738,936,937,939,942,945,948],{"class":740,"line":516},[738,938,767],{"class":750},[738,940,941],{"class":754}," config",[738,943,944],{"class":754}," set",[738,946,947],{"class":754}," ai.provider",[738,949,950],{"class":754}," \"ollama\"\n",[20,952,212],{"id":212},[214,954,955,971],{},[217,956,957],{},[220,958,959,961,963,965,968],{},[223,960,225],{},[223,962,566],{},[223,964,230],{},[223,966,967],{},"AutoGLM",[223,969,970],{},"Crush",[238,972,973,989,1004,1018,1034,1048,1063,1078,1094,1110],{},[220,974,975,977,980,983,986],{},[243,976,245],{},[243,978,979],{},"IM 网关 + 多 Agent 自治",[243,981,982],{},"桌面客户端 + MCP",[243,984,985],{},"Phone\u002FWeb GUI agent",[243,987,988],{},"终端 TUI agent",[220,990,991,994,997,1000,1002],{},[243,992,993],{},"7×24 Heartbeat",[243,995,996],{},"✅ 30 分钟",[243,998,999],{},"–",[243,1001,999],{},[243,1003,999],{},[220,1005,1006,1009,1012,1014,1016],{},[243,1007,1008],{},"多 IM 渠道",[243,1010,1011],{},"✅ 20+",[243,1013,999],{},[243,1015,999],{},[243,1017,999],{},[220,1019,1020,1023,1026,1029,1031],{},[243,1021,1022],{},"Skills 市场",[243,1024,1025],{},"✅ ClawHub 5400+",[243,1027,1028],{},"✅ MCP Connectors",[243,1030,999],{},[243,1032,1033],{},"✅ Agent Skills 标准",[220,1035,1036,1038,1040,1043,1045],{},[243,1037,600],{},[243,1039,295],{},[243,1041,1042],{},"Projects",[243,1044,999],{},[243,1046,1047],{},"session",[220,1049,1050,1052,1054,1057,1060],{},[243,1051,261],{},[243,1053,273],{},[243,1055,1056],{},"闭源",[243,1058,1059],{},"✅ MIT 模型 + Apache 代码",[243,1061,1062],{},"FSL-1.1-MIT",[220,1064,1065,1068,1071,1073,1076],{},[243,1066,1067],{},"本地 LLM",[243,1069,1070],{},"✅ Ollama 友好",[243,1072,999],{},[243,1074,1075],{},"✅ 9B 自模型",[243,1077,295],{},[220,1079,1080,1083,1086,1089,1091],{},[243,1081,1082],{},"中国大陆友好",[243,1084,1085],{},"✅ 完全",[243,1087,1088],{},"❌ 支付",[243,1090,295],{},[243,1092,1093],{},"✅ 部分",[220,1095,1096,1099,1102,1105,1107],{},[243,1097,1098],{},"起价",[243,1100,1101],{},"$0",[243,1103,1104],{},"$20\u002F月 +",[243,1106,1101],{},[243,1108,1109],{},"$0 + API",[220,1111,1112,1115,1118,1121,1124],{},[243,1113,1114],{},"适合",[243,1116,1117],{},"7×24 IM 触发个人 agent",[243,1119,1120],{},"MCP 生态桌面",[243,1122,1123],{},"Phone GUI",[243,1125,1126],{},"终端 power user",[20,1128,350],{"id":350},[40,1130,1131,1137,1143,1149,1155,1161,1167,1173,1179,1185],{},[43,1132,1133,1136],{},[46,1134,1135],{},"Skill 选型评估","：5400+ Skill 里挑官方 \u002F 高 star \u002F 仔细看 source，沙箱模式优先",[43,1138,1139,1142],{},[46,1140,1141],{},"API 预算控制","：Heartbeat + 多 Agent 会放大 LLM 费用，配 budget cap 或本地 Ollama",[43,1144,1145,1148],{},[46,1146,1147],{},"国内 IM 风控","：微信 \u002F 钉钉账号要用独立小号，避免主账号被风控",[43,1150,1151,1154],{},[46,1152,1153],{},"memory 膨胀","：长期记忆 JSON 文件定期归档 \u002F 清理",[43,1156,1157,1160],{},[46,1158,1159],{},"Node 版本","：Node 24 推荐，Node 18 部分新 Skill 不兼容",[43,1162,1163,1166],{},[46,1164,1165],{},"sandbox 权限默认保守","：复杂自动化要在 onboard 阶段调整权限，不要一上来全 root",[43,1168,1169,1172],{},[46,1170,1171],{},"数据备份","：~\u002F.openclaw 含所有记忆 + 配置，定期备份",[43,1174,1175,1178],{},[46,1176,1177],{},"Windows 走 WSL2","：原生 Windows 兼容差",[43,1180,1181,1184],{},[46,1182,1183],{},"MCP 复用","：内部 MCP server 可直接调用，不要重复造轮",[43,1186,1187,1190],{},[46,1188,1189],{},"Heartbeat 间隔自定义","：默认 30 分钟，做高频监控可调短，但成本对应增加",[20,1192,405],{"id":404},[40,1194,1195,1198,1201,1204,1207,1210,1213,1216],{},[43,1196,1197],{},"✅ 想从 IM 一句话支使 AI 7×24 干活的 power user",[43,1199,1200],{},"✅ 监控 \u002F 定时简报 \u002F 邮件分拣 \u002F 行情盯盘 \u002F 智能家居",[43,1202,1203],{},"✅ 数据隐私敏感 + 自托管刚需",[43,1205,1206],{},"✅ 中国大陆用户走本地 Ollama 全免费",[43,1208,1209],{},"❌ 纯 GUI \u002F 桌面交互党（用 Claude Desktop）",[43,1211,1212],{},"❌ 不会折腾 Node \u002F Docker \u002F WSL",[43,1214,1215],{},"❌ 要一键就跑的非技术用户",[43,1217,1218],{},"❌ 高强度多 Agent 同时跑 + 预算极紧（API 费会膨胀）",[20,1220,469],{"id":469},[40,1222,1223,1229,1235],{},[43,1224,1225],{},[473,1226,1228],{"href":1227},"\u002Ftools\u002Fagent\u002Fdesktop\u002Fclaude-desktop","Claude Desktop 评测",[43,1230,1231],{},[473,1232,1234],{"href":1233},"\u002Ftools\u002Fagent\u002Fdesktop\u002Fautoglm","AutoGLM 评测",[43,1236,1237],{},[473,1238,1240],{"href":1239},"\u002Ftools\u002Fagent\u002Fdesktop\u002Fcrush","Crush 评测",[20,1242,488],{"id":488},[173,1244,1245,1252,1259,1266],{},[43,1246,1247,1248],{},"OpenClaw 官方文档（架构 \u002F Gateway \u002F 渠道 \u002F Heartbeat）",[473,1249,1250],{"href":1250,"rel":1251},"https:\u002F\u002Fdocs.openclaw.ai\u002F",[502],[43,1253,1254,1255],{},"技术栈 — OpenClaw 2026 实战指南 ",[473,1256,1257],{"href":1257,"rel":1258},"https:\u002F\u002Fjishuzhan.net\u002Farticle\u002F2058753885218062337",[502],[43,1260,1261,1262],{},"CSDN — OpenClaw 完全指南（35.5 万 stars \u002F Heartbeat \u002F 多 Agent）",[473,1263,1264],{"href":1264,"rel":1265},"https:\u002F\u002Fblog.csdn.net\u002Fweixin_58240054\u002Farticle\u002Fdetails\u002F161197395",[502],[43,1267,1268,1269],{},"腾讯云 — OpenClaw 爆火 + Computer Use Agent 分析 ",[473,1270,1271],{"href":1271,"rel":1272},"https:\u002F\u002Fcloud.tencent.com\u002Fdeveloper\u002Farticle\u002F2637863",[502],[1274,1275,1276],"style",{},"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 pre.shiki code .sVt8B, html code.shiki .sVt8B{--shiki-default:#24292E;--shiki-dark:#E1E4E8}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":512,"searchDepth":513,"depth":513,"links":1278},[1279,1280,1281,1282,1283,1284,1285,1286,1287,1288],{"id":22,"depth":516,"text":23},{"id":38,"depth":516,"text":38},{"id":98,"depth":516,"text":98},{"id":668,"depth":516,"text":669},{"id":171,"depth":516,"text":171},{"id":212,"depth":516,"text":212},{"id":350,"depth":516,"text":350},{"id":404,"depth":516,"text":405},{"id":469,"depth":516,"text":469},{"id":488,"depth":516,"text":488},"\u002Fimg\u002Ftools\u002Fopenclaw.webp","OpenClaw 真实评测：原名 Clawdbot \u002F Moltbot，Peter Steinberger 2025 末创建，2026-02 移交开源基金会。60 天 28-35.5 万 GitHub stars，超 React 历史速度。MIT 协议自托管 Gateway 连接 Discord \u002F Slack \u002F Telegram \u002F WhatsApp \u002F iMessage \u002F Matrix \u002F Signal \u002F Microsoft Teams \u002F Zalo 等 20+ IM 平台。Heartbeat 30 分钟主动触发 + Cron + Skills（ClawHub 5400+）+ 多 Agent 隔离 + 本地长期记忆。Node 24（推荐）或 22 LTS。",[1292,1295,1298,1301],{"q":1293,"a":1294},"OpenClaw 和 Claude Desktop \u002F MCP 怎么定位？","Claude Desktop 是『桌面客户端 + MCP 工具栈』走 GUI 交互；OpenClaw 是『IM 网关 + 多 Agent + 主动 Heartbeat』走聊天 + 后台自治。两者互补：在电脑前要 GUI 体验用 Claude Desktop；离开电脑要 Telegram \u002F 微信里支使 AI 7×24 干活用 OpenClaw。OpenClaw 内部也可调用 MCP server 和 LLM provider。",{"q":1296,"a":1297},"Heartbeat 是什么？","每 30 分钟由 Gateway 主动唤醒 Agent，检查 Cron \u002F 任务队列 \u002F 监控钩子，让 AI 真正『7×24 后台运行』而非被动等用户。监控告警 \u002F 定时简报 \u002F 邮件分拣 \u002F 行情盯盘等场景靠 Heartbeat 实现。",{"q":1299,"a":1300},"ClawHub 是什么？","OpenClaw 社区驱动的 Skills 市场，2026-05 已突破 5400 个 Skill：天气 \u002F 邮件分拣 \u002F 网页采集 \u002F 浏览器自动化 \u002F 文件整理 \u002F 多平台消息 \u002F 加密货币 \u002F 智能家居 \u002F CRM 集成等。Skill 用 YAML \u002F Markdown 描述，一键安装 + 自定义 prompt 即可。",{"q":1302,"a":1303},"中国大陆能用吗？","完全可以。OpenClaw 自托管 + MIT 开源 + 本地 Ollama 模型路径全免费 + 无外网 \u002F 海外卡依赖。微信 \u002F 钉钉 \u002F 飞书都有 channel 插件。配 Ollama llama3 \u002F Qwen 中文模型可全离线。仅外接 GPT \u002F Claude API 时需要海外网络 + 海外卡。",[533,1305,1306],"zh","multi",{},"\u002Ftools\u002Fagent\u002Fdesktop\u002Fopenclaw",[541,542,1310,786],"windows-wsl",[1312],{"plan":658,"price":1101,"features":1313,"notes":1314},"MIT 全开源 + Gateway + Skills + Memory + Cron + Heartbeat + 多 Agent + 20+ IM 渠道","用户自带 LLM Key 或本地 Ollama","MIT 完全免费开源 \u002F 用户自带 LLM API（GPT \u002F Claude \u002F 本地 Ollama 全免费）","2026-06-19",[1318],"onboarding\u002Fpersonal-247-ai-agent",{"power":547,"ux":546,"price":547,"cn_support":546,"stability":546},{"title":566,"description":1290},"OpenClaw - 开源个人 AI Agent 评测与指南 | AIHO",[1323,1326,1328,1330],{"name":1324,"url":1250,"accessed":1325},"OpenClaw 官方文档（架构 \u002F Gateway \u002F 多渠道）","2026-06-24",{"name":1327,"url":1257,"accessed":1325},"技术栈 — OpenClaw 2026 实战指南",{"name":1329,"url":1264,"accessed":1325},"CSDN — OpenClaw 完全指南（35.5 万 stars \u002F Heartbeat）",{"name":1331,"url":1271,"accessed":1325},"腾讯云 — Computer Use Agent 爆发分析","tools\u002Fagent\u002Fdesktop\u002Fopenclaw","2026 现象级开源个人 AI Agent——多 IM 网关 + 技能市场 ClawHub 5400+ + 7×24 Heartbeat",[558,1335,1336,1337,1338,1339,1340,767],"self-host","gateway","mcp","multi-agent","im-bot","mit","2026 最火个人 AI Agent——『IM 触发 + 7×24 主动 + 本地控制』三件套。要从 Telegram \u002F 微信 \u002F Discord 一句话支使 AI 干活的 power user 首选。GUI 党 \u002F 非工程师仍推 Claude Desktop。","https:\u002F\u002Fdocs.openclaw.ai","liK3OxB1DF-JGg4dNwI6i7wgA6iE8mkxCT63ALFupPA",1785428440463]