Data Intelligence MCP Server

How to use

You connect MCP clients to the Data Intelligence MCP Server either locally in stdio mode or remotely via HTTP/HTTPS. In stdio mode, the client invokes the server directly through standard input/output, which is ideal for local development and testing. In HTTP/HTTPS mode, you run the server as a standalone service and configure clients to reach it over HTTP(S) using the provided API keys or credentials.

For local setups, use stdio mode to simplify integration with Claude Desktop, VS Code Copilot, and other MCP clients. Configure each client with a dedicated MCP server entry that points to the local server command and environment variables. When remote operation is required, you can run the HTTP endpoint and provide the necessary authentication headers or tokens in your client configuration.

How to install

Prerequisites: Python 3.11 or higher and Data Intelligence SaaS or CPD 5.2.1.

1) Install the MCP server package.

pip install ibm-watsonx-data-intelligence-mcp-server

2) (Optional) Install the UV runtime helper if you plan to run in stdio mode with uvx.

# Example if you need uvx tooling; follow your environment's guidance for installing uvx
# This step assumes uvx is available on your system

3) Run the server in stdio mode for local development.

uvx ibm-watsonx-data-intelligence-mcp-server --transport stdio

4) (Optional) Run the server in HTTP mode if you have a remote or local HTTP(S) endpoint you want to expose.

ibm-watsonx-data-intelligence-mcp-server --transport http --host 0.0.0.0 --port 3000

Configuration and notes

The server can be configured via environment variables or a .env file. Copy .env.example to .env and modify values as needed.

Key client environment variables you will use (examples shown; replace with your actual values):

DI_SERVICE_URL=https://api.dataplatform.cloud.ibm.com
DI_ENV_MODE=SaaS
REQUEST_TIMEOUT_S=60
LOG_FILE_PATH=/tmp/di-mcp-server-logs
DI_CONTEXT=df

Client configuration

Configure your MCP client to connect to the local stdio server or to the HTTP(S) endpoint. The following configurations illustrate both local stdio connections and remote HTTP setups.

Claude Desktop (stdio, recommended for local setup) requires a small JSON snippet to point to the local server.

{
  "mcpServers": {
    "wxdi-mcp-server": {
      "command": "uvx",
      "args": ["ibm-watsonx-data-intelligence-mcp-server", "--transport", "stdio"],
      "env": {
         "DI_SERVICE_URL": "https://api.dataplatform.cloud.ibm.com",
         "DI_APIKEY": "<data intelligence api key>",
         "DI_ENV_MODE": "SaaS",
         "LOG_FILE_PATH": "/tmp/di-mcp-server-logs"
      }
    }
  }
}

SSL/TLS Configuration

If you are running in CPD or a secured environment, you may need a TLS/SSL configuration for the client connection. Look into SSL certificate requirements and headers when configuring remote HTTP(S) connections.

Security and access

Use strong API keys or Bearer tokens for remote HTTP(S) connections. For local stdio mode, ensure the environment and filesystem permissions protect sensitive logs and data.

Notes

The Data Intelligence MCP Server is designed to integrate with IBM Data Intelligence services and supports modular interaction patterns with MCP clients. When deploying in production, prefer HTTP(S) mode with proper security and access controls, and ensure environment variables are securely managed.