MCP management
AI analysis often requires connecting to external resources, such as fetching deal data from a CRM system or pushing generated analysis reports to an instant messaging group. A Model Context Protocol (MCP) connector integrates data and capabilities from external systems into your AI analysis workflow, expanding the AI's capabilities. On the MCP management page, administrators can view, manage, and add connectors for their organization. This document explains how to add and manage these connectors.
Prerequisites
Subscription: Your organization must have an active Quick BI AIPro subscription.
Permissions: Only an administrator with centralized management permissions for Quick BI AIPro can access the Management Center to add and manage MCP connectors. For more information about configuring permissions, see User Permissions.
Access the feature
Navigate to the Management Center in Quick BI AIPro and click MCP management in the left navigation pane.

Add an MCP connector
On the MCP management page, click Add Connector in the upper-right corner.

In the Import by JSON dialog box, enter the configuration information for the connector.

Parameter
Description
① Configure JSON
Enter the configuration JSON for the connector. Two modes are supported: STDIO and SSE.
STDIO: The client and server communicate directly through standard input/output streams without a network port. This is suitable for scenarios such as local development, debugging, and standalone deployment. Supported startup commands include
npxanduvx. The following is an example configuration:// stdio example: { "mcpServers": { "stdio-server-example": { "command": "npx", "args": ["-y", "mcp-server-example"] } } }NoteWhen adding an MCP that uses
npxto start, the system performs a pre-check and displays a notification: "Detected that this MCP is launched using npx, which may prolong the startup time for subsequent sessions." This notification does not prevent you from adding the connector.SSE: The server pushes event messages to the client over a persistent HTTP connection. This mode is ideal for scenarios that require server-pushed data, remote deployment, or shared services across multiple clients. The following is an example configuration:
// SSE example: { "mcpServers": { "sse-server-example": { "url": "http://localhost:8080/sse" } } }
② Connector description
Enter a description for the connector, such as its function, business purpose, or usage instructions.
Click Import to add the MCP connector. By default, the connector is enabled and visible to all users in your organization after import. Users can then select it in the task dialog box when creating new analysis tasks.

Manage MCP connectors
The MCP management page lists all MCP connectors added to your organization, along with their description, visibility scope, creator, and status. From this page, you can perform the following management operations.

Actions | Description |
① Search for a connector | If you have many connectors, enter keywords in the search box to quickly find a specific one. |
② Set status | A connector is enabled by default after being added. Use the toggle in the Enabled column to change its status. |
③ Edit | Click the |
④ Set visibility scope | Click the
|
⑤ Delete | Click the |
⑧ Batch operation | You can select and delete multiple connectors at once. |
icon to open the Edit Knowledge dialog box to modify the connector's configuration information.
