Deploy Precisely MCP server - Precisely Data Integrity Suite

1. Download and install VS Code from https://code.visualstudio.com.
2. Install the GitHub Copilot extension.
   1. Press Ctrl+Shift+X.
   2. Search GitHub Copilot.
   3. Select Install.
3. Sign in to GitHub Copilot when prompted.
4. Create a new workspace folder called Suite MCP in your C: drive.
5. Go to File > Open Folder and open Suite MCP.
6. Log into your Data Integrity Suite workspace at https://cloud.precisely.com.
7. Select Account.
8. Select the API Keys page.
9. Select Generate API Key or select an existing API key.
10. Copy the generated credential value for VS Code.
11. In your workspace folder, create a .github folder to add the MCP configuration file
12. Open the context menu for the .github folder and create a file named mcp.json.
13. Paste the following content into mcp.json:

     {
      "servers": {
        "dis-mcp": {
          "type": "http",
          "url": "<MCP_SERVER_URL>",
          "headers": {
                    "Authorization": "Apikey <generated_credential_value>"
          }
        }
      }
    }

14. Replace generated_credential_value with the credential value copied from the API Keys page.
15. Save the file.

Verify the connection

16. Press Ctrl+Shift+P and type MCP: List Servers.
17. Confirm that DIS-MCP is listed with status Running.
    Tip: If it shows an error, copy the generated credential value from the API Keys page again and update the MCP server configuration.

1. Go to Customize > Connectors.
2. Select + then Add custom connector.
3. Enter the following details:
   • Name: Suite MCP
   • For Remote MCP server URL, enter <MCP_SERVER_URL>.
4. Select Advanced settings and enter the following OAuth credentials:

   Field	Value
   OAuth Client ID	0oawtlk3vhx09KMUJ4x7
   OAuth Client Secret	Leave blank (public client)

5. Select Add.
6. Select Connect and complete the Precisely SSO login.
7. For Team and Enterprise accounts: Only an Owner can add a custom connector for the organization. Members skip straight to this step. Go to Customize > Connectors, find Precisely Suite MCP (marked Custom), and select Connect.

1. Open the config file for your operating system:

   Operating System	Path
   macOS	~/Library/Application Support/Claude/claude_desktop_config.json
   Windows	%APPDATA%\Claude\claude_desktop_config.json

2. Merge the following configuration into the mcpServers object. Replace generated_credential_value with the credential value copied for Claude Desktop from your Data Integrity Suite workspace.

   Windows:

   {
     "mcpServers": {
       "Suite MCP": {
         "command": "cmd",
         "args": ["/c", "npx", "-y", "mcp-remote",
           "<MCP_SERVER_URL>",
                   "--header", "Authorization: Apikey <generated_credential_value>"]
       }
     }
   }

   macOS / Linux:

   {
     "mcpServers": {
       "Suite MCP": {
         "command": "npx",
         "args": ["-y", "mcp-remote",
           "<MCP_SERVER_URL>",
                   "--header", "Authorization: Apikey <generated_credential_value>"]
       }
     }
   }

3. Verify the header format is strict: Authorization:Apikey <value> (no extra space after the colon).
4. Quit and relaunch Claude Desktop. Precisely Suite MCP appears in Customize > Connectors.
5. Once connected, enable the connector per conversation via the + button → Connectors → toggle Precisely Suite MCP on.

1. Create the Unity Catalog connection.
   1. Go to Catalog.
   2. Select + then Create a connection.
   3. Set the following fields and select Next:
      • Connection name: precisely_dis_mcp
      • Connection type: HTTP
      • Auth type: OAuth Machine to Machine
2. Fill in the OAuth credentials and select Next:

   Field	Value
   Host	<MCP_SERVER_HOST>
   Port	Keep defaults
   Client ID	API key created in your Suite workspace (see Create an API Key)
   Client secret	API secret created in your Suite workspace (see Create an API Key)
   Token endpoint	https://api.cloud.precisely.com/auth/v2/token
   OAuth Scope	default

3. Configure the MCP-specific settings:
   1. Check Is mcp connection.
   2. Set Base path to /mcp.
4. Select Create connection.
5. Create an AI Agent backed by the precisely_dis_mcp connection.
   1. In the left navigation, go to ML or AI → Agents.
   2. Select Create Agent.
   3. Select Supervisor Agent.
   4. Under Tools and sub-agents, select Add an External MCP and select precisely_dis_mcp.
   5. Set Instructions. Example:

      You help users enrich and validate data using the Suite. Use precisely_actions_search to discover relevant actions, precisely_actions_describe to understand their inputs, and precisely_actions_execute to run them.

   6. Select Open in Playground.
6. Test the agent in AI Playground. Follow Validate connectivity via chat and try the example queries. The agent automatically selects and calls the appropriate Suite MCP tool based on your request.

1. Open Microsoft Copilot Studio.
2. Select Agents to create a new agent
3. Choose Create blank agent.
4. Provide the agent name, description, and instructions:
   • Name (for example, "Suite MCP Assistant")
   • Description (for example, "An agent that searches, describes, and runs Precisely Suite actions via MCP")
   • Instructions - Example: "You help users discover and run Precisely data intelligence actions. Use the Suite MCP tools to search for available actions, describe their inputs and outputs, and run them on your behalf."
5. Select your agent's model (for example, GPT-4o or the default model available in your environment).
6. Save the agent.
7. Open your agent to enable generative orchestration.
8. Go to Settings.
9. Ensure Generative orchestration is turned On.
10. Open your agent again to add the Suite MCP server as a tool.
11. Go to the Tools page.
12. Select Add a tool→ Create New→Model Context Protocol.

    This launches the MCP onboarding wizard.

13. To configure the MCP server connection, Provide the following details in the wizard:

    Field	Value
    Server name	Precisely Suite MCP
    Description	Precisely Suite MCP server - search, describe, and run data intelligence actions
    Server URL	<MCP_SERVER_URL>

    Authentication: Select API Key

    • Type: Header
    • Header name: Authorization
14. Select Create to create the MCP connection.
15. On the Add Tools page, confirm that Connection shows Not connected.
16. Select the dropdown next to it and select Create new connection.
17. When prompted, enter the credential value copied for Microsoft Copilot Studio from the API Keys page.
18. Select Create.
19. Confirm the new connection is visible.
20. Select Add and Configure. The tool appears on the Tools page.
21. After adding the MCP server, confirm that the following three tools are available to the agent:

    Tool	Purpose
    precisely_actions_search	Discover actions by natural language goal
    precisely_actions_describe	Get full schema and examples for a specific action
    precisely_actions_execute	Run an action with provided arguments

    Tool changes on the MCP server are dynamically reflected in Copilot Studio.

22. Open Test your agent in Copilot Studio and try sample queries.

    The agent automatically selects and calls the appropriate Suite MCP tool based on what you ask.

23. Select Publish.
24. Choose a channel (for example, Microsoft Teams, Microsoft 365).
25. Configure visibility and submit for approval if required.

1. Test Discovery: Search for actions related to geocoding an address.

   You should see it invoke the precisely_actions_search tool and return a list of matching actions.

2. Test Detail: Describe the geocode.address action.

   You should see it invoke the precisely_actions_describe tool and return the action's schema and examples.

3. Test: What's the flood risk for 1600 Pennsylvania Ave NW, Washington, DC?

   You should see it invoke precisely_actions_execute and return a result. An authentication error means your credential value is invalid or lacks permission to perform actions. Copy the generated credential value from the API Keys page again, then update your configuration.