Skip to main content
Glama
Sign In
YingDao RPA MCP Server

YingDao RPA MCP Server

Official
by ying-dao
OverviewSchemaSourceRelated ServersDiscussionsAdmin
TypeScript
Hybrid

影刀RPA MCP Server

影刀RPA:一个RPA低代码平台，一款人人可用的RPA自动化产品，能够将人从重复的劳动中解放出来。 影刀AI Power:一个AI低代码平台，能够快速创建AI智能体、AI工作流，帮助用户把AI用起来。

影刀 RPA MCP Server 基于 Model Context Protocol (MCP) 实现，为影刀AI Power及其他可作为MCP Host的工具(如 Claude Desktop、Cursor 等)提供调用RPA的能力。

同时支持SSE Server与Stdio Server两种模式。

如何开始

支持两种方式来运行影刀RPA

本地模式

设置环境变量 注意：本地模式下，智能获取并运行“我获取的应用”并已经被至少执行过一次的应用

RPA_MODEL=local
SHADOWBOT_PATH={your_shadowbot_path} //影刀rpa的exe路径
USER_FOLDER={your_user_folder}       //影刀rpa的用户文件夹路径

影刀RPA的exe路径

Windows 注意：windows下在AI Power客户端中，路径要使用双斜杠

D://Program Files//{安装目录}//ShadowBot.exe

Mac

/Applications/影刀.app

影刀RPA的用户文件夹路径

在影刀RPA的设置中，找用户文件夹选项

Related MCP server: Integrator MCP Server

开放API模式 (仅支持企业用户)

设置环境变量

RPA_MODEL=openApi
ACCESS_KEY_ID={your_access_key_id}
ACCESS_KEY_SECRET={your_access_key_secret}

获取方式

企业管理员登录影刀RPA控制台获取，请参考影刀RPA帮助文档-鉴权

Stdio Server启动

在客户端中配置

{
  "mcpServers": {
    "YingDao RPA MCP Server": {
      "command": "npx",
      "args": ["-y", "yingdao-mcp-server"],
      "env":{
        "RPA_MODEL":"openApi",
        "ACCESS_KEY_ID":"{your_access_key_id}",
        "ACCESS_KEY_SECRET":"{your_access_key_secret}"
      }
    }
  }
}

SSE Server配置

构建

Clone the repository and build:

git clone https://github.com/ying-dao/yingdao_mcp_server.git
cd yingdao_mcp_server
npm install
npm run build

配置

添加.env文件，配置项参考以上描述

启动

npm run start:server

客户端配置

AI Power 客户端配置

{
  "mcpServers": {
    "YingDao RPA MCP Server": {
      "url": "http://localhost:3000/sse",
      "description": "影刀 MCP Server"
    }
  }
}

默认端口为3000

能力

本地模式

  1. queryRobotParam: 查询RPA应用的参数

  2. queryApplist: 查询RPA应用的列表

  3. runApp: 运行RPA应用

开放API模式

  1. uploadFile: 上传文件到RPA平台

  2. queryRobotParam: 查询RPA应用参数

  3. queryApplist: 获取分页RPA应用列表

  4. startJob: 启动RPA作业

  5. queryJob: 查询RPA作业状态

  6. queryClientList: 查询调度RPA机器人列表

License

MIT

Available Tools

6 tools
queryApplistB

该接口用于分页获取RPA应用列表。

Inputs: appId, size, page, ownerUserSearchKey, +1 more.

queryClientListC

该接口用于查询RPA机器人列表。

Inputs: status, key, robotClientGroupUuid, page, +1 more.

queryJobA

该接口用于查询RPA应用JOB状态。

Inputs: jobUuid (required).

queryRobotParamC

该接口用于查询RPA机器人参数。

Inputs: robotUuid, accurateRobotName.

startJobC

该接口用于启动RPA应用JOB。

Inputs: robotUuid (required), accountName, robotClientGroupUuid, waitTimeoutSeconds, +2 more.

uploadFileC

该接口用于上传文件到RPA平台。

Inputs: fileName (required), file.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    Deploy Server
    One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

    TDQS

    B3.2/5.0

    Scored across 6 tools

    Disambiguation5/5

    Each tool targets a distinct resource or action: app list, client list, job status, robot parameters, job start, and file upload. There is no clear overlap between tool purposes.

    Naming Consistency5/5

    All tool names follow a consistent verb+noun camelCase pattern (e.g., queryApplist, startJob, uploadFile). The verbs are descriptive and the pattern is uniform.

    Tool Count5/5

    Six tools is a reasonable scope for an RPA server, covering essential queries, job execution, and file upload without being too sparse or overwhelming.

    Completeness3/5

    The set covers querying and starting jobs but lacks CRUD operations for apps/clients and job stop functionality. Notable gaps exist for a full lifecycle.

    Maintenance

    ActivityInactive
    ResponsivenessLimited data

    Resources

    Related MCP Connectors

    Related MCP Servers