Data Product

Log in to Data Mesh Manager, navigate to your account settings, and generate an API key. This is required for all tool calls.

Depending on which output ports you want to query, collect the appropriate credentials: Snowflake username/password/warehouse/role, Databricks client ID/secret/host/HTTP path, or a BigQuery service account JSON path.

Edit your claude_desktop_config.json (or equivalent) to register the dataproduct_mcp server with uvx and all required environment variables.

{
  "mcpServers": {
    "dataproduct": {
      "command": "uvx",
      "args": ["dataproduct_mcp"],
      "env": {
        "DATAMESH_MANAGER_API_KEY": "your-api-key",
        "DATACONTRACT_SNOWFLAKE_USERNAME": "your-username",
        "DATACONTRACT_SNOWFLAKE_PASSWORD": "your-password",
        "DATACONTRACT_SNOWFLAKE_WAREHOUSE": "your-warehouse",
        "DATACONTRACT_SNOWFLAKE_ROLE": "your-role"
      }
    }
  }
}

Restart Claude Desktop (or your chosen MCP client) to load the new server configuration. The dataproduct server should appear in the connected servers list.

Use the dataproduct_search tool to find data products by keyword. Then use dataproduct_get to retrieve full details including output ports and schema information.

Call dataproduct_request_access with a data_product_id, output_port_id, and business purpose. Once access is granted, use dataproduct_query to execute SQL and retrieve up to 100 rows.