[{"_1":2,"_337":-5,"_338":-5},"loaderData",{"_3":4,"_53":54,"_55":56,"_250":251},"root",{"_5":6,"_49":50,"_51":52},"appGlobal",{"_7":8,"_38":39,"_40":41},"config",{"_9":10,"_11":12,"_13":14,"_15":16,"_17":18,"_19":20,"_21":22,"_23":-5,"_24":25,"_26":27,"_28":29,"_30":31,"_32":33,"_34":31,"_35":31,"_36":37},"APP_URL","https://glama.ai","CLIENT_ASSETS_URL","https://static.glama.ai/client/","ENVIRONMENT","LIVE","FACEBOOK_PIXEL_ID","1988603181578092","GOOGLE_ANALYTICS_MEASUREMENT_ID","G-GDQ5NW5CB5","GOOGLE_OAUTH2_CLIENT_ID","48173796263-7ohlqirg28gh2a7nel61coff3m1ogjtc.apps.googleusercontent.com","INTERCOM_APP_ID","wgcqevfz","INTERCOM_USER_HASH","POSTHOG_API_KEY","phc_aK0RnJxBEBRE0h0LIz21yyc9w6ylBxwUHMPB4712Bjc","REDDIT_PIXEL_ID","a2_fcqsoymvje51","RELEASE_VERSION","glama@production","SENTRY_DRY_RUN",false,"SENTRY_DSN","https://fc82aa9f808b13927ff29c56bda4d486@o4507975513735168.ingest.us.sentry.io/4507975514849280","SENTRY_SPOTLIGHT","SENTRY_TRACE","STRIPE_PUBLIC_KEY","pk_live_51QpZQ32LV3sZcHqzHlFuNhgSfggSNaQqCzZw9EMJ8tbhpQyZgcwQj1vz168FxQ35EGRcztKu2MankxiSs7zLCF4g00blUverGs","viewportWidth",1280,"visitorSession",{"_42":43,"_45":-5,"_46":47,"_48":-5},"attributes",[44],"BOT","membership","uid","we6jmudana","userAccount","nonce","f1f93a78f3d6e48392c52605f1a45b67","toasts",[],"routes/_public/_layout",{},"routes/_public/mcp/servers/~uid/_pages/_layout",{"_57":-5,"_58":59,"_169":170},"githubUser","mcpServer",{"_60":61,"_62":63,"_42":70,"_73":74,"_75":74,"_76":77,"_82":83,"_105":106,"_107":108,"_101":109,"_110":-5,"_111":-5,"_112":113,"_119":-5,"_120":121,"_126":-5,"_127":128,"_132":133,"_46":166,"_167":168},"addedAt",["D",1735550975887],"argumentsJsonSchema",{"_64":65,"_66":67,"_68":69},"type","object","required",[],"properties",{},[71,72],"category:home-automation-and-iot","language:python","descriptionMarkdown","Links IoT devices to AI large models using the MCP and MQTT protocols, enabling natural language control, real-time AI responses, and complex instruction execution for interconnected IoT devices.","descriptionPlainText","githubProject",{"_78":79,"_80":81},"readme","# mcp2mqtt: 连接物理世界与AI大模型的桥梁 \n\n[English](README_EN.md) | 简体中文\n\n

\n $\"mcp2mqtt$ \n

通过自然语言控制硬件，开启物联网新纪元

\n\n## 系统架构\n\n

\n $\"系统架构图\"$ \n

mcp2mqtt 系统架构图

\n\n## 工作流程\n\n

\n $\"工作流程图\"$ \n

mcp2mqtt 工作流程图

\n\n## 项目愿景\n\nmcp2mqtt 是一个将物联网设备接入AI大模型的项目，它通过 Model Context Protocol (MCP) 和 MQTT 协议将物理世界与 AI 大模型无缝连接。最终实现：\n- 用自然语言控制你的硬件设备\n- AI 实时响应并调整物理参数\n- 让你的设备具备理解和执行复杂指令的能力\n- 通过MQTT协议实现设备间的互联互通\n\n## 主要特性\n\n- **智能MQTT通信**\n - 支持MQTT协议的发布/订阅模式\n - 支持多种MQTT服务器（如Mosquitto、EMQ X等）\n - 支持QoS服务质量保证\n - 支持主题过滤和消息路由\n - 实时状态监控和错误处理\n\n- **MCP 协议集成**\n - 完整支持 Model Context Protocol\n - 支持资源管理和工具调用\n - 灵活的提示词系统\n - 通过MQTT实现命令的发布与响应\n\n## 配置说明\n\n### MQTT配置\n```yaml\nmqtt:\n broker: \"localhost\" # MQTT服务器地址\n port: 1883 # MQTT服务器端口\n client_id: \"mcp2mqtt_client\" # MQTT客户端ID\n username: \"mqtt_user\" # MQTT用户名\n password: \"mqtt_password\" # MQTT密码\n keepalive: 60 # 保持连接时间\n topics:\n command:\n publish: \"mcp/command\" # 发送命令的主题\n subscribe: \"mcp/response\" # 接收响应的主题\n status:\n publish: \"mcp/status\" # 发送状态的主题\n subscribe: \"mcp/control\" # 接收控制命令的主题\n```\n\n### 命令配置\n```yaml\ncommands:\n set_pwm:\n command: \"CMD_PWM {frequency}\"\n need_parse: false\n data_type: \"ascii\"\n prompts:\n - \"把PWM调到最大\"\n - \"把PWM调到最小\"\n mqtt_topic: \"mcp/pwm\" # MQTT发布主题\n response_topic: \"mcp/pwm/response\" # MQTT响应主题\n```\n\n## MQTT 命令和响应\n\n### 命令格式\n\n命令使用简单的文本格式：\n\n1. PWM 控制：\n - 命令：`PWM {值}`\n - 示例：\n - `PWM 100`（最大值）\n - `PWM 0`（关闭）\n - `PWM 50`（50%）\n - 响应：`CMD PWM {值} OK`\n\n2. LED 控制：\n - 命令：`LED {状态}`\n - 示例：\n - `LED on`（打开）\n - `LED off`（关闭）\n - 响应：`CMD LED {状态} OK`\n\n3. 设备信息：\n - 命令：`INFO`\n - 响应：`CMD INFO {设备信息}`\n\n### 错误响应\n\n如果发生错误，响应格式将为：\n`ERROR: {错误信息}`\n\n## 支持的客户端\n\nmcp2mqtt 支持所有实现了 MCP 协议的客户端，以及支持MQTT协议的物联网设备：\n\n| 客户端类型 | 特性支持 | 说明 |\n|--------|----------|------|\n| Claude Desktop | 完整支持 | 推荐使用，支持所有 MCP 功能 |\n| Continue | 完整支持 | 优秀的开发工具集成 |\n| Cline | 资源+工具 | 支持多种 AI 提供商 |\n| MQTT设备 | 发布/订阅 | 支持所有MQTT协议的物联网设备 |\n\n## 快速开始\n\n### 1. 安装\n\n#### Windows用户\n下载 [install.py](https://raw.githubusercontent.com/mcp2everything/mcp2mqtt/main/install.py) \n```bash\npython install.py\n```\n#### macOS用户\n```bash\n# 下载安装脚本\ncurl -O https://raw.githubusercontent.com/mcp2everything/mcp2mqtt/main/install_macos.py\n\n# 运行安装脚本\npython3 install_macos.py\n```\n\n#### Ubuntu/Raspberry Pi用户\n```bash\n# 下载安装脚本\ncurl -O https://raw.githubusercontent.com/mcp2everything/mcp2mqtt/main/install_ubuntu.py\n\n# 运行安装脚本\npython3 install_ubuntu.py\n```\n\n安装脚本会自动完成以下操作：\n- 检查系统环境\n- 安装必要的依赖\n- 创建默认配置文件\n- 配置Claude桌面版（如果已安装）\n\n### 手动分步安装依赖\n```bash\nwindows\npowershell -ExecutionPolicy ByPass -c \"irm https://astral.sh/uv/install.ps1 | iex\"\nMacOS\ncurl -LsSf https://astral.sh/uv/install.sh | sh\n```\n主要依赖uv工具，所以当python和uv以及Claude或Cline安装好后就可以了。\n\n### 基本配置\n在你的 MCP 客户端（如 Claude Desktop 或 Cline）配置文件中添加以下内容：\n注意：如果使用的自动安装那么会自动配置Calude Desktop无需此步。\n使用默认配置文件：\n```json\n{\n \"mcpServers\": {\n \"mcp2mqtt\": {\n \"command\": \"uvx\",\n \"args\": [\n \"mcp2mqtt\"\n ]\n }\n }\n}\n```\n> 注意：修改配置后需要重启Cline或者Claude客户端软件\n## 配置说明\n### 配置文件位置\n复制配置文件（`config.yaml`）可以放在位置：\n用户主目录（推荐个人使用）\n```bash\n# Windows系统\nC:\\Users\\用户名\\.mcp2mqtt\\config.yaml\n\n# macOS系统\n/Users/用户名/.mcp2mqtt/config.yaml\n\n# Linux系统\n/home/用户名/.mcp2mqtt/config.yaml\n```\n- 适用场景：个人配置\n- 需要创建 `.mcp2mqtt` 目录：\n ```bash\n # Windows系统（在命令提示符中）\n mkdir \"%USERPROFILE%\\.mcp2mqtt\"\n \n # macOS/Linux系统\n mkdir -p ~/.mcp2mqtt\n ```\n \n指定配置文件：\n比如指定加载Pico配置文件：Pico_config.yaml\n```json\n{\n \"mcpServers\": {\n \"mcp2mqtt\": {\n \"command\": \"uvx\",\n \"args\": [\n \"mcp2mqtt\",\n \"--config\",\n \"Pico\" //指定配置文件名，不需要添加_config.yaml后缀\n ]\n }\n }\n}\n```\n为了能使用多个mqtt，我们可以新增多个mcp2mqtt的服务指定不同的配置文件名即可。\n如果要接入多个设备，如有要连接第二个设备：\n指定加载Pico2配置文件：Pico2_config.yaml\n```json\n{\n \"mcpServers\": {\n \"mcp2mqtt2\": {\n \"command\": \"uvx\",\n \"args\": [\n \"mcp2mqtt\",\n \"--config\",\n \"Pico2\" //指定配置文件名，不需要添加_config.yaml后缀\n ]\n }\n }\n}\n```\n\n### 硬件连接\n1. 将你的设备通过网络连接到mqtt服务器\n2. 也可以用tests目录下的responder.py来模拟设备\n\n## 运行测试\n\n### 启动设备模拟器\n\n项目在 `tests` 目录中包含了一个设备模拟器。它可以模拟一个硬件设备，能够：\n- 响应 PWM 控制命令\n- 提供设备信息\n- 控制 LED 状态\n\n启动模拟器：\n```bash\npython tests/responder.py\n```\n\n你应该能看到模拟器正在运行并已连接到 MQTT 服务器的输出信息。\n\n### 启动客户端Claude 桌面版或Cline\n

\n $\"Cline$ \n

Example in Cline

\n\n### 从源码快速开始\n1. 从源码安装\n```bash\n# 通过源码安装：\ngit clone https://github.com/mcp2everything/mcp2mqtt.git\ncd mcp2mqtt\n\n# 创建虚拟环境\nuv venv .venv\n\n# 激活虚拟环境\n# Windows:\n.venv\\Scripts\\activate\n# Linux/macOS:\nsource .venv/bin/activate\n\n# 安装开发依赖\nuv pip install --editable .\n```\n\n### MCP客户端配置\n在使用支持MCP协议的客户端（如Claude Desktop或Cline）时，需要在客户端的配置文件中添加以下内容：\n直接自动安装的配置方式\n源码开发的配置方式\n#### 使用默认演示参数：\n```json\n{\n \"mcpServers\": {\n \"mcp2mqtt\": {\n \"command\": \"uv\",\n \"args\": [\n \"--directory\",\n \"你的实际路径/mcp2mqtt\", // 例如: \"C:/Users/Administrator/Documents/develop/my-mcp-server/mcp2mqtt\"\n \"run\",\n \"mcp2mqtt\"\n ]\n }\n }\n}\n```\n#### 指定参数文件名\n```json\n{\n \"mcpServers\": {\n \"mcp2mqtt\": {\n \"command\": \"uv\",\n \"args\": [\n \"--directory\",\n \"你的实际路径/mcp2mqtt\", // 例如: \"C:/Users/Administrator/Documents/develop/my-mcp-server/mcp2mqtt\"\n \"run\",\n \"mcp2mqtt\",\n \"--config\", // 可选参数，指定配置文件名\n \"Pico\" // 可选参数，指定配置文件名，不需要添加_config.yaml后缀\n ]\n }\n }\n}\n```\n

\n $\"Cline$ \n

Example in Cline

\n### 配置文件位置\n配置文件（`config.yaml`）可以放在不同位置，程序会按以下顺序查找：\n#### 1. 当前工作目录（适合开发测试）\n- 路径：`./config.yaml`\n- 示例：如果你在 `C:\\Projects` 运行程序，它会查找 `C:\\Projects\\config.yaml`\n- 适用场景：开发和测试\n- 不需要特殊权限\n\n#### 2. 用户主目录（推荐个人使用）\n```bash\n# Windows系统\nC:\\Users\\用户名\\.mcp2mqtt\\config.yaml\n\n# macOS系统\n/Users/用户名/.mcp2mqtt/config.yaml\n\n# Linux系统\n/home/用户名/.mcp2mqtt/config.yaml\n```\n- 适用场景：个人配置\n- 需要创建 `.mcp2mqtt` 目录：\n ```bash\n # Windows系统（在命令提示符中）\n mkdir \"%USERPROFILE%\\.mcp2mqtt\"\n \n # macOS/Linux系统\n mkdir -p ~/.mcp2mqtt\n ```\n\n#### 3. 系统级配置（适合多用户环境）\n```bash\n# Windows系统（需要管理员权限）\nC:\\ProgramData\\mcp2mqtt\\config.yaml\n\n# macOS/Linux系统（需要root权限）\n/etc/mcp2mqtt/config.yaml\n```\n- 适用场景：多用户共享配置\n- 创建目录并设置权限：\n ```bash\n # Windows系统（以管理员身份运行）\n mkdir \"C:\\ProgramData\\mcp2mqtt\"\n \n # macOS/Linux系统（以root身份运行）\n sudo mkdir -p /etc/mcp2mqtt\n sudo chown root:root /etc/mcp2mqtt\n sudo chmod 755 /etc/mcp2mqtt\n ```\n\n程序会按照上述顺序查找配置文件，使用找到的第一个有效配置文件。根据你的需求选择合适的位置：\n- 开发测试：使用当前目录\n- 个人使用：建议使用用户主目录（推荐）\n- 多用户环境：使用系统级配置（ProgramData或/etc）\n\n3. 运行服务器：\n```bash\n# 确保已激活虚拟环境\n.venv\\Scripts\\activate\n\n# 运行服务器（使用默认配置config.yaml 案例中用的LOOP_BACK 模拟串口，无需真实串口和串口设备）\nuv run src/mcp2mqtt/server.py\n或\nuv run mcp2mqtt\n# 运行服务器（使用指定配置Pico_config.yaml）\nuv run src/mcp2mqtt/server.py --config Pico\n或\nuv run mcp2mqtt --config Pico\n```\n\n\n## 文档\n\n- [安装指南](./docs/zh/installation.md)\n- [API文档](./docs/zh/api.md)\n- [配置说明](./docs/zh/configuration.md)","url","https://github.com/mcp2everything/mcp2mqtt","githubRepository",{"_84":85,"_86":87,"_88":89,"_90":91,"_92":93,"_97":98,"_99":100,"_103":104},"defaultBranch","main","fullName","mcp2everything/mcp2mqtt","githubTreeId",1406,"language","Python","owner",{"_94":95,"_96":31},"login","mcp2everything","verified","recentStargazers",1,"spdxLicense",{"_101":102},"name","MIT License","stargazers",7,"id",505,"inspectable",true,"mcp2mqtt","npmPackage","prompts","relatedMcpServerIds",[114,115,116,117,118],491,315,361,494,323,"resources","scores",{"_122":123,"_124":123,"_125":123},"license",100,"quality","security","simpleIcon","supportedPlatforms",[129,130,131],"WINDOWS","MACOS","LINUX","tools",[134,146,152],{"_101":135,"_136":137,"_138":139},"set_pwm","description","设置PWM频率，范围0-100","inputSchema",{"_64":65,"_66":140,"_68":142},[141],"frequency",{"_141":143},{"_64":144,"_136":145},"integer","PWM频率值(0-100)",{"_101":147,"_136":148,"_138":149},"get_pico_info","获取Pico开发板信息",{"_64":65,"_66":150,"_68":151},[],{},{"_101":153,"_136":154,"_138":155},"led_control","控制LED开关",{"_64":65,"_66":156,"_68":158},[157],"state",{"_157":159},{"_160":161,"_64":164,"_136":165},"enum",[162,163],"on","off","string","LED状态(on/off)","bunlaxlfyz","updatedAt",["D",1740226794333],"meta",[171,174,176,183,186,188,191,194,197,200,203,205,208,211,213,215,217,219,221],{"_172":173},"title","mcp2mqtt | Glama",{"_175":74,"_101":136},"content",{"_177":178,"_179":180,"_181":182},"href","https://glama.ai/mcp/servers/bunlaxlfyz","rel","canonical","tagName","link",{"_175":109,"_184":185},"property","og:title",{"_175":74,"_184":187},"og:description",{"_175":189,"_184":190},"https://glama.ai/logo.png","og:logo",{"_175":192,"_184":193},"https://glama.ai/mcp/servers/bunlaxlfyz/og-image","og:image",{"_175":195,"_184":196},"630","og:image:height",{"_175":198,"_184":199},"1200","og:image:width",{"_175":201,"_184":202},"website","og:type",{"_175":178,"_184":204},"og:url",{"_175":206,"_184":207},"ChatGPT alternative for power users","og:site_name",{"_175":209,"_101":210},"summary_large_image","twitter:card",{"_175":206,"_101":212},"twitter:title",{"_175":74,"_101":214},"twitter:description",{"_175":192,"_101":216},"twitter:image",{"_175":195,"_101":218},"twitter:image:height",{"_175":198,"_101":220},"twitter:image:width",{"_222":223},"script:ld+json",{"_224":225,"_226":227},"@context","http://schema.org","@graph",[228],{"_224":229,"_230":231,"_232":233},"https://schema.org","@type","BreadcrumbList","itemListElement",[234,240,244,248],{"_230":235,"_236":237,"_101":238,"_239":98},"ListItem","item","https://glama.ai/","Glama","position",{"_230":235,"_236":241,"_101":242,"_239":243},"https://glama.ai/mcp","MCP",2,{"_230":235,"_236":245,"_101":246,"_239":247},"https://glama.ai/mcp/servers","Servers",3,{"_230":235,"_236":178,"_101":109,"_239":249},4,"routes/_public/mcp/servers/~uid/_pages/tools/~name/_route",{"_58":252,"_282":-5,"_283":284,"_169":291,"_328":329,"_40":336},{"_60":253,"_62":254,"_42":257,"_73":74,"_75":74,"_76":258,"_82":259,"_105":106,"_107":108,"_101":109,"_110":-5,"_111":-5,"_112":262,"_119":-5,"_120":263,"_126":-5,"_127":264,"_132":265,"_46":166,"_167":281},["D",1735550975887],{"_64":65,"_66":255,"_68":256},[],{},[71,72],{"_78":79,"_80":81},{"_84":85,"_86":87,"_88":89,"_90":91,"_92":260,"_97":98,"_99":261,"_103":104},{"_94":95,"_96":31},{"_101":102},[114,115,116,117,118],{"_122":123,"_124":123,"_125":123},[129,130,131],[266,271,275],{"_101":135,"_136":137,"_138":267},{"_64":65,"_66":268,"_68":269},[141],{"_141":270},{"_64":144,"_136":145},{"_101":147,"_136":148,"_138":272},{"_64":65,"_66":273,"_68":274},[],{},{"_101":153,"_136":154,"_138":276},{"_64":65,"_66":277,"_68":278},[157],{"_157":279},{"_160":280,"_64":164,"_136":165},[162,163],["D",1740226794333],"mcpServerInstance","mcpServerTool",{"_136":154,"_105":285,"_138":286,"_101":153},1067,{"_68":287,"_66":290,"_64":65},{"_157":288},{"_136":165,"_160":289,"_64":164},[162,163],[157],[292,294,295,296,298,299,300,301,302,303,304,305,306,307,308,309,310,311,312],{"_172":293},"led_control - mcp2mqtt | Glama",{"_175":154,"_101":136},{"_177":178,"_179":180,"_181":182},{"_175":297,"_184":185},"led_control - mcp2mqtt",{"_175":154,"_184":187},{"_175":189,"_184":190},{"_175":192,"_184":193},{"_175":195,"_184":196},{"_175":198,"_184":199},{"_175":201,"_184":202},{"_175":178,"_184":204},{"_175":206,"_184":207},{"_175":209,"_101":210},{"_175":206,"_101":212},{"_175":154,"_101":214},{"_175":192,"_101":216},{"_175":195,"_101":218},{"_175":198,"_101":220},{"_222":313},{"_224":225,"_226":314},[315],{"_224":229,"_230":231,"_232":316},[317,318,319,320,321,325],{"_230":235,"_236":237,"_101":238,"_239":98},{"_230":235,"_236":241,"_101":242,"_239":243},{"_230":235,"_236":245,"_101":246,"_239":247},{"_230":235,"_236":178,"_101":109,"_239":249},{"_230":235,"_236":322,"_101":323,"_239":324},"https://glama.ai/mcp/servers/bunlaxlfyz/tools","Tools",5,{"_230":235,"_236":326,"_101":153,"_239":327},"https://glama.ai/mcp/servers/bunlaxlfyz/tools/led_control",6,"otherTools",[330,332,334],{"_58":331,"_101":147},{"_46":166},{"_58":333,"_101":153},{"_46":166},{"_58":335,"_101":135},{"_46":166},{"_48":-5},"actionData","errors"]

led_control

Input Schema

Input Schema (JSON Schema)

Other Tools