modelscope·@ying-dao/yingdao_mcp_server
一种模型上下文协议服务器,使人工智能平台能够调用英道的RPA(机器人流程自动化)功能,通过人工智能交互实现重复性任务的自动执行。
影刀RPA: An RPA low-code platform, a product that everyone can use for RPA automation, which can free people from repetitive labor. <br/> 影刀AI Power: An AI low-code platform that allows for the quick creation of AI agents and AI workflows, helping users to utilize AI.
The YingDao RPA MCP Server is implemented based on the Model Context Protocol (MCP), providing the capability to call RPA for YingDao AI Power and other tools that can act as MCP Hosts (such as Claude Desktop, Cursor, etc.).
It supports both SSE Server and Stdio Server modes.
There are two ways to run YingDao RPA:
Set environment variables Note: In local mode, only "My Acquired Apps" that have been executed at least once can be intelligently retrieved and run. bash RPA_MODEL=local SHADOWBOT_PATH={your_shadowbot_path} // Path to the YingDao RPA .exe file USER_FOLDER={your_user_folder} // Path to the YingDao RPA user folder
Windows Note: On Windows, in the AI Power client, use double backslashes for the path. bash D:\Program Files\{installation_directory}\ShadowBot.exe
Mac
bash /Applications/影刀.app
Find the user folder option in the settings of YingDao RPA.
<p><img src="src/assets/user_folder.png" width="512"/></p>Set environment variables
bash RPA_MODEL=openApi ACCESS_KEY_ID={your_access_key_id} ACCESS_KEY_SECRET={your_access_key_secret}
Enterprise administrators can log into the YingDao RPA console to obtain these. Please refer to YingDao RPA Help Documentation - Authentication.
Configure in the client json { "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}" } } } }
Clone the repository and build:
bash git clone https://github.com/ying-dao/yingdao_mcp_server.git cd yingdao_mcp_server npm install npm run build
Add an .env file, with configuration items as described above.
bash npm run start:server
AI Power client configuration json { "mcpServers": { "YingDao RPA MCP Server": { "url": "http://localhost:3000/sse", "description": "YingDao MCP Server" } } }
The default port is 3000.
MIT