# icedq-tools/mcp-server [Health: Active]

**Category:** 📊 Data Platforms  
**Repository:** https://github.com/icedq-tools/mcp-server  
**GitHub Stars:** 1  
**npm Downloads (last month):** 232  
**Views:** 2  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/mcp-server-124

## Description
MCP Server for iceDQ Data Reliability Platform

## Claude Desktop Quick Installation
Install path detected from listing signals. Uses `npx` (confidence: high):

```json
"mcpServers": {
  "mcp-server": {
    "command": "npx",
    "args": ["-y","@icedq/mcp-server"],
    "env": {
      "ICEDQ_BASE_URL": "",
      "ICEDQ_REALM": "",
      "ICEDQ_CLIENT_ID": "",
      "AUTH_TYPE": "",
      "ICEDQ_ORG_ID": "",
      "ICEDQ_CLIENT_SECRET": "",
      "ICEDQ_USERNAME": "",
      "ICEDQ_PASSWORD": ""
    }
  }
}
```

**Requires environment variables:** `ICEDQ_BASE_URL`, `ICEDQ_REALM`, `ICEDQ_CLIENT_ID`, `AUTH_TYPE`, `ICEDQ_ORG_ID`, `ICEDQ_CLIENT_SECRET`, `ICEDQ_USERNAME`, `ICEDQ_PASSWORD` — the values above are empty placeholders; fill in real credentials before running (see the repository for what each one is for).

## Documentation

## What icedq-tools/mcp-server does

icedq-tools/mcp-server MCP server gives AI assistants an MCP interface to the iceDQ Data Reliability Platform. It is intended for teams that want conversational access to data-quality resources and operations in an iceDQ instance.

The server covers the main stages of a data-quality workflow. An assistant can browse workspaces, connections, databases, schemas, tables, columns, files, and REST APIs. It can retrieve sample data, inspect quality indicators such as nulls, patterns, and types, and request check recommendations based on a profile.

Created rules can validate rows with NotNull, Format, ValidValues, Length, Date, or custom Groovy checks. Other supported rule types handle duplicate detection, SQL-based pushdown checks, cross-source checksums, and row-level reconciliation across databases, files, and APIs. Reconciliation includes AI-assisted join-key mapping. Rules created through the MCP tools are published immediately.

## How it works

The server is launched by an MCP-compatible client and authenticates against an iceDQ instance using configured environment variables.

After rules are created, the assistant can execute them on demand, inspect status, and view exception reports. Multiple rules can be grouped into sequential workflows. Schedules support one-time, daily, or weekly execution. The server also includes eight agent skills that provide guidance for tasks such as authoring, comparison, API testing, runs, and schedules.

## Setup and configuration

icedq-tools/mcp-server MCP server is published as the `@icedq/mcp-server` npm package and can normally be started with `npx -y @icedq/mcp-server`. The client configuration supplies the command, package argument, and environment variables.

A working setup requires an iceDQ 7.5.0 or newer instance, a valid user account, an organization ID, and authentication details. Required configuration includes the iceDQ base URL, authentication realm, OAuth client ID, authentication type, and organization ID. Username-password authentication additionally uses the client secret, username, and password. Access-token authentication uses a token JSON file, while device flow opens a browser login and caches tokens in the operating system credential store.

Node.js 20 or newer is required for the `npx` installation path. Claude Desktop's packaged extension includes its own runtime. Windows 10 or newer, macOS 10.15 or newer, and Linux are supported; Linux users must use the `npx` approach because the packaged Claude Desktop extension is limited to macOS and Windows.

## Tools and capabilities

- Explore iceDQ workspaces, connections, schemas, tables, columns, files, and APIs.
- Profile sample data and inspect null rates, patterns, and data types.
- Create validation, duplicate, pushdown, checksum, and reconciliation rules.
- Build workflows, schedules, folders, reusable parameters, and Java/Groovy user-defined functions.
- Run checks, monitor execution, and inspect exception reports.
- Query structured datasets containing iceDQ execution history.

## Limitations and notes

The server depends on access to an iceDQ environment and does not operate as a standalone data-quality engine. Its documented compatibility is with Claude Desktop, VS Code, Cursor, and Claude Code. Configuration must match the selected authentication mode, and credentials or token files must be available to the process.

Rules created through the MCP interface are published immediately, so users should verify parameters and targets before authoring them.

_Full upstream README: https://allmcps.com/mcp/mcp-server-124/readme_