Overview

The iceDQ MCP Server connects AI assistants — including Claude Desktop, VS Code, and Cursor — to your iceDQ Data Quality Platform, enabling intelligent data quality management through natural language.

With 48 tools across the full data quality lifecycle, you can explore data landscapes, profile quality metrics, generate AI-powered check suggestions, create and execute validation rules, and monitor results — all from within your AI assistant.

Installation

The server is published to npm as @icedq/mcp-server. Most MCP clients can launch it on demand with npx — no manual install or build step required.

Add this to your MCP client configuration (e.g. Claude Desktop's claude_desktop_config.json, VS Code's .vscode/mcp.json, or Cursor's MCP settings):

{
  "mcpServers": {
    "icedq": {
      "command": "npx",
      "args": ["-y", "@icedq/mcp-server"],
      "env": {
        "ICEDQ_BASE_URL": "https://app.icedq.net",
        "ICEDQ_REALM": "iam.icedq",
        "ICEDQ_CLIENT_ID": "<your-client-id>",
        "ICEDQ_CLIENT_SECRET": "<your-client-secret>",
        "AUTH_TYPE": "username_password",
        "ICEDQ_USERNAME": "<your-username>",
        "ICEDQ_PASSWORD": "<your-password>",
        "ICEDQ_ORG_ID": "<your-org-id>"
      }
    }
  }
}

Prefer a global install? npm install -g @icedq/mcp-server installs the icedq-mcp-server command, which your client can run directly instead of npx.

Environment Variables

| Variable | Required | Description | |----------|----------|-------------| | ICEDQ_BASE_URL | ✅ | Base URL of your iceDQ instance (e.g. https://app.icedq.net) | | ICEDQ_REALM | ✅ | Authentication realm (default iam.icedq) | | ICEDQ_CLIENT_ID | ✅ | OAuth client ID for API authentication | | AUTH_TYPE | ✅ | username_password or access_token | | ICEDQ_ORG_ID | ✅ | Your iceDQ organization ID | | ICEDQ_CLIENT_SECRET | ⬩ | Required for username_password mode | | ICEDQ_USERNAME | ⬩ | Required for username_password mode | | ICEDQ_PASSWORD | ⬩ | Required for username_password mode | | TOKENS_PATH | ⬩ | Path to a token JSON file — required for access_token mode | | DEBUG | — | Set true for verbose logging |

For per-client walkthroughs and authentication details, see the setup guides below.

Getting Started

Choose your setup guide based on your AI client:

| Client | Setup Guide | |--------|-------------| | Claude Desktop | Installation in Claude Desktop | | VS Code | VS Code & Cursor Setup | | Cursor | VS Code & Cursor Setup | | Authentication | Authentication Guide |

For a general introduction, see the MCP Server Overview.

Tools Reference

48 Tools Across 10 Categories

| Category | Tools | |----------|-------| | Discovery | list_workspaces, list_connections, test_connection, list_folders, list_rules, list_workflows, list_schedules | | Data Exploration | list_databases, list_schemas, list_tables, list_columns, get_database_metadata, fetch_sample_data | | Profiling & AI | profile_data, suggest_quality_checks | | Validation Rules | create_validation_rule, update_rule, get_rule | | Duplicate Detection | create_duplicate_rule | | SQL Pushdown Rules | create_pushdown_rule | | Cross-Source Rules | create_checksum_rule, analyze_recon_mapping, create_recon_rule | | Workflows & Schedules | create_workflow, add_rules_to_workflow, remove_rules_from_workflow, create_schedule, modify_schedule, add_rules_workflows_to_schedule | | Execution & Monitoring | execute_rule, execute_schedule, check_workflow_run_status, get_workflow_run_result, get_checks_exception_report, get_exception_report_url, get_rule_workflow_run_history, get_scheduler_runs_history | | Organization | create_folder, move_rules, move_workflows, check_task_status, create_parameter, update_parameter, parse_csv_and_create_parameter, get_guidance |

Data Warehouse Queries

| Tool | Description | |------|-------------| | datawarehouse_query_schema | Get the data warehouse query schema definition | | datawarehouse_query_executor | Execute structured queries against the iceDQ data warehouse | | validate_and_explain_structured | Validate and preview a query before execution |

Usage Examples

Discover Your Data Landscape

"List my iceDQ workspaces and show the connections in the first one"

Profile Data and Get Quality Check Suggestions

"Fetch sample data from dbo.Customers, profile it, and suggest quality checks"

Create a Validation Rule

"Create a validation rule on dbo.Customers with NotNull on Email, Email format check, and ValidValues on Status"

Cross-Source Checksum Comparison

"Compare row counts between SQL Server Orders and Snowflake FACT_ORDERS"

Execute and Monitor a Rule

"Execute the Customer_Quality_Checks rule and show me the exception report"

Troubleshooting

| Issue | Solution | |-------|----------| | Organization ID required | Set ICEDQ_ORG_ID in your MCP configuration | | Authentication failures | Verify ICEDQ_CLIENT_ID, ICEDQ_CLIENT_SECRET, username, and password | | No workspaces returned | Check base URL, verify credentials, and confirm the user has workspace access | | Token file errors | Ensure TOKENS_PATH points to a valid, readable JSON file with accessToken and refreshToken fields | | Sample data not returning | Confirm the connection is ACTIVE and the table name is correct (case-sensitive) |

To enable detailed logging, set DEBUG=true in your MCP environment configuration.

Security

• Credentials are stored and managed by your MCP client — never by the MCP server
• All communication uses HTTPS with OAuth 2.0 token-based authentication
• In access_token mode, tokens are automatically refreshed and persisted to the configured token file (TOKENS_PATH)
• No data is transmitted to third-party services or telemetry systems

Requirements

• Node.js 18.0.0 or higher
• iceDQ 7.5.0+ with a valid user account
• MCP Client: Claude Desktop, VS Code, or Cursor (latest versions)
• OS: Windows 10+, macOS 10.15+, or Linux

Support

• Documentation: https://docs.icedq.com/guides/mcp-server/intro
• Email: [email protected]
• Website: https://icedq.com
• License: MIT