Configuring watsonx Orchestrate
Register watsonx.data intelligence MCP server in the watsonx Orchestrate configuration to complete tasks across data governance and catalog, data quality, data lineage, and Data Product Hub with the AI agent.
Required configuration information
- Authentication: API key; configure the client to send your watsonx platform API key on each request. For more information, see Generating API keys and Available APIs.
- Server URL: The service endpoint is based on your IBM Software Hub deployment URL (
{cpd_cluster_host}). Example:https://{cpd_cluster_host}/semantic_agents/public/v1/mcp_server/mcp/. For more information, see Service endpoint in the API documentation. - Transport:
http
Configuring watsonx Orchestrate
-
If you don't have a connection in watsonx Orchestrate that you can use to connect with the watsonx.data intelligence service, create one:
- Go to Manage > Connections and select Add new connection.
- Define the connection details.
- Select your authentication type.
If you select API key, before you can use the tools, you must connect to your instance by using your API key. The connection is remembered unless you reset it. - Specify your credential type.
- Click Connect.
- Click Add connection.
-
Build and deploy a new agent:
- Go to Build > Agent Builder and click Create Agent.
- Select Create from scratch.
- Type in a name and description.
- Click Create.
-
Add the remote IBM watsonx.data intelligence MCP server to your toolset:
- In Agent chat > Manage agents > Your_agent, open Toolset.
- Click Add tool.
- Click Create a new flow > Agentic workflow > MCP server.
- Click Add MCP server.
- For Select an MCP server type, select Remote MCP server.
- Specify the server name, MCP server URL, and connection.
-
Select the remote IBM watsonx.data intelligence MCP server tool. In Toolset, click Add tool to select a tool.
When you're working with your AI agent, you can select tools to help complete your tasks. Limit your selection to a subset of tools which you consider useful.
The agent is now ready to use. Open Chat from the navigation menu and select your agent from the drop-down menu to start chatting with it.