ecs_mcp_server

This MCP server enables AI models to manage 360 AI 云盘 (cloud disk) files through the MCP protocol. It exposes a variety of file management capabilities, including listing, searching, uploading, downloading, saving from URLs or text, creating folders, sharing files, moving and renaming items, and retrieving user information. Typical workflows let an AI agent query the cloud disk contents, upload assets from a prompt, or save content directly into the云盘, with downloads and share links supplied back to the user. The server supports multiple interaction modes; primarily the Stdio interface via MCP Server (the npx-based setup), but there are also HTTP-based streams available for integration without local Node.js, and SSE for real-time server push scenarios.

Prerequisites:

• Node.js (LTS) and npm installed on your machine or server
• Access/permissions to run npx and access to the internet to fetch the MCP server package

Install and run via Stdio MCP (recommended):

1. Ensure Node.js and npm are installed

   • node -v
   • npm -v

2. Run the MCP server using npx as configured in mcp_config

   • Create or edit your Cursor config (as shown in the README), or run directly: npx -y @aicloud360/mcp-server-disk

3. Provide your API key when prompted or set it in the environment variable API_KEY

   • On Unix-like systems: export API_KEY=_xxxxxxxxx
   • On Windows (PowerShell): $env:API_KEY = "_xxxxxxxxx"

4. Verify the server starts and is reachable at the expected endpoint defined by your client setup (Stdio MCP via Cursor, or via the provided URL for HTTP/SSE modes).

Alternative access modes (no local Node.js required):

• Streamable HTTP: configure the HTTP URL in Cursor using the provided endpoint and API key
• SSE: configure the SSE URL in Cursor for real-time streaming

Notes:

• The npm package name used in the config is @aicloud360/mcp-server-disk; using npx will fetch and run it without a local install.
• Replace _xxxxxxxxx with a valid API key obtained from the 360 AI 云盘开放平台.

Tips and common issues:

• Ensure the API_KEY is valid and has the required permissions for 云盘 operations.
• If you encounter network or authentication errors, verify that the API key is correctly passed in the env or URL parameter as documented for the chosen connection method.
• The HTTP Streamable and SSE modes do not require a local Node.js environment but rely on accessible endpoints. Use them if you want to integrate MCP with web-based clients or services without Node.js runtime.
• Supported MCP actions include: file-list, file-search, file-upload-stdio, file-download-stdio, file-save, video-download, make-dir, file-share, file-move, file-rename, user-info.
• When using file-upload-stdio or file-download-stdio, ensure the client provides proper paths and permissions for reading local files or writing downloads.
• For large video downloads via video-download, consider increasing client timeout/timeout_settings to accommodate longer transfers.