qiniu-mcp-server is a Model Context Protocol (MCP) Server built on Qiniu Cloud products. Its goal is to let users access Qiniu Cloud Storage, Intelligent Multimedia, Live Streaming, and other capabilities through this MCP Server in the context of an AI large model client.
Purpose This service is used to connect Qiniu Cloud's product capabilities to AI large model clients that support MCP, so that AI can directly operate Qiniu Cloud-related resources during a conversation.
Main capabilities Storage supports: getting the Bucket list, getting the file list in a Bucket, uploading local files as well as uploading by providing file content, reading file content, and getting file download links. Intelligent Multimedia supports: image scaling and image rounded corners. CDN supports: refreshing files by link and prefetching files by link. Live Streaming supports: creating a live streaming space bucket, creating a live stream, getting the live streaming space list, getting the stream list, binding push/pull stream domains, getting push/pull live streaming addresses, and getting live streaming usage.
How to use Environment requirements: Python 3.12 or later; if uv is not installed yet, you can install it with the corresponding command (the specific installation command content is not provided in the materials). Download the Cline plugin in vscode. After downloading, a Cline icon will be added to the sidebar. Configure qiniu MCP: click the Cline icon to enter the Cline plugin, select the MCP Server module, select installed, click Advanced MCP Settings to configure the MCP Server. Refer to the following configuration information: { "mcpServers": { "qiniu": { "command": "uvx", "args": [ "qiniu-mcp-server" ], "env": { "QINIU_ACCESS_KEY": "YOUR_ACCESS_KEY", "QINIU_SECRET_KEY": "YOUR_SECRET_KEY", "QINIU_REGION_NAME": "YOUR_REGION_NAME", "QINIU_ENDPOINT_URL": "YOUR_ENDPOINT_URL", "QINIU_BUCKETS": "YOUR_BUCKET_A,YOUR_BUCKET_B" }, "disabled": false } } } If only using the live streaming feature, two authentication modes are supported: 1 Configure QINIU_ACCESS_KEY/QINIU_SECRET_KEY: { "mcpServers": { "qiniu": { "command": "uvx", "args": [ "qiniu-mcp-server" ], "env": { "QINIU_ACCESS_KEY": "YOUR_ACCESS_KEY", "QINIU_SECRET_KEY": "YOUR_SECRET_KEY" }, "disabled": false } } } 2 After obtaining the apikey in the Qiniu Live Streaming console, configure QINIU_LIVE_API_KEY: { "mcpServers": { "qiniu": { "command": "uvx", "args": [ "qiniu-mcp-server" ], "env": { "QINIU_LIVE_API_KEY": "YOUR_LIVE_API_KEY" }, "disabled": false } } } Click the link switch of the qiniu MCP Server to connect.
Create a chat window in Cline, and you can interact with AI to use qiniu-mcp-server. Examples for object storage are: list all Buckets in qiniu; list the files in the xxx Bucket in qiniu; read the file content of yyy in the qiniu xxx Bucket; apply a rounded corner with a width of 200 pixels to the image yyy in the qiniu xxx Bucket; refresh this CDN link of qiniu:https://developer.qiniu.com/test.txt。 Examples of the live streaming system: list all live streaming spaces; create 1 new live streaming space named mcptest1123; create 1 new stream for mcptest1123 named stream1; list all streams in the live streaming space mcptest1123; bind a push domain and a pull domain to mcptest1123; obtain the push URL and pull URL corresponding to the live stream stream1 under mcptest1123; obtain the usage of live streaming in the last 1 hour.
Applicable scenarios and limitations The MCP Server is suitable for accessing and managing Qiniu Cloud storage, intelligent multimedia, CDN, and live streaming-related resources in AI large model clients that support MCP. The materials mention that creating an MCP Server in cursor can directly use the above configuration; when using it in claude, you may encounter: Error: spawn uvx ENOENT error. Solution: fill in the absolute path of uvx in the command parameter, eg: /usr/local/bin/uvx.
Extension and testing Extension function: first add a new business package directory under the core directory (eg: storage -> storage), and complete the function extension in this business package directory; in the init.py file under the business package directory, define the load function to register business tools or resources, and finally call this load function in the init.py under the core directory to complete the registration of tools or resources. Testing: It is strongly recommended to use Model Control Protocol Inspector for testing. Example of starting MCP Server locally: start in standard input/output (stdio) mode (default); start in SSE mode (for web applications). Edit the .env file and configure the following parameters (the materials do not list the specific parameter contents).
