Cloud control MCP
What is MCP
tip
- MCP (Model Context Protocol) is an open protocol from Anthropic that standardizes how LLMs talk to external data and tools — like a “USB-C for AI”
- Overview: https://baike.baidu.com/item/%E6%A8%A1%E5%9E%8B%E4%B8%8A%E4%B8%8B%E6%96%87%E5%8D%8F%E8%AE%AE/65540618
- Cloud control speaks MCP, so you can drive it with natural language via an LLM — no custom code required
Supported platforms
How to use cloud MCP
- Install cloud control. MCP listens on port
8187— open it in BT Panel security and your cloud provider security group - Go to Permission → User management → Cloud settings
- If there is no MCP key, generate one, copy it, then confirm to save
MCP server config
- After you have the key, edit the config below:
- Replace the server IP with your cloud control IP
- Replace the MCP key with the key you copied
{
"mcpServers": {
"EcloudMCP_zuhu": {
"type": "streamable-http",
"url": "http://SERVER_IP:8187/mcp",
"headers": {
"UserToken": "MCP_KEY"
}
}
}
}
Example:
{
"mcpServers": {
"EcloudMCP_zuhu": {
"type": "streamable-http",
"url": "http://192.168.2.12:8187/mcp",
"headers": {
"UserToken": "6c2657067fb7459bb7f84b6afab5e5b1"
}
}
}
}
Configure in an LLM platform
- Example uses Trae; other platforms are similar
- Download and open Trae
- Open AI feature management on the right
- Open MCP → Add → Manual
- Paste the MCP config above and confirm
- Trae connects and lists supported tools
- Click Agent → create
- Name the agent, check the MCP server you configured, then create
- Chat with the agent, e.g.
Run cloud task named 323orStop cloud task named 323
Why MCP helps
- Often used in workflows as a tool/data provider for the LLM
- Example: Coze generates a video, then starts a cloud task for the next step — one of many scenarios
- For simple automation, skip coding and let the model interpret intent
Feature requests
- For more MCP ideas or needs, contact Secretary Sha (QQ 2557945562)