Get the FREE Ultimate OpenClaw Setup Guide →

mdm

mdm-mcp-server is a Model Context Protocol (MCP) server that provides AI assistants with seamless access to IBM Master Data Management (MDM) services through standardized REST API interactions.

Installation
Run this command in your terminal to add the MCP server to Claude Code.
Run in terminal:
Command
claude mcp add --transport stdio ibm-mdm-mcp-server uvx ibm_mdm_mcp_server --mode stdio \
  --env API_CLOUD_CRN="<your_instance_crn>" \
  --env MCP_TOOLS_MODE="minimal" \
  --env API_CLOUD_API_KEY="<your_api_key>" \
  --env API_CLOUD_AUTH_URL="https://iam.cloud.ibm.com/identity/token" \
  --env API_CLOUD_BASE_URL="<your_mdm_base_url>" \
  --env M360_TARGET_PLATFORM="cloud"

How to use

This IBM MDM MCP Server provides a standardized bridge between AI assistants and IBM Master Data Management (MDM) services. It exposes a RESTful interface compatible with the MCP (Model Context Protocol) so tools in Claude Desktop or other MCP clients can discover and invoke MDM capabilities like searching master data, retrieving data models, and manipulating records. The server supports two tool exposure modes: minimal (essential tools only) and full (all available tools). When integrated with Claude Desktop, you can configure the MCP server as a uvx-based source, enabling STDIO-style communication that Claude can use for natural language querying and tool invocation. Typical workflows include asking for a data model, querying records, or performing lookups and entity resolutions via the MCP tools.

How to install

Prerequisites:

  • Python 3.10 or newer
  • Git (optional for cloning)
  • Access to an IBM MDM instance (Cloud or Software Hub) with credentials

Option 1: Install from PyPI (Recommended)

pip install ibm-mdm-mcp-server

After installation, run the server directly (HTTP mode by default):

ibm_mdm_mcp_server

Option 2: Quick Start with Setup Wizard

# Clone the repository (optional if you use PyPI, but provided for setup convenience)
git clone https://github.com/IBM/mdm-mcp-server.git
cd mdm-mcp-server

# Run the setup wizard
python setup_wizard.py

The wizard will guide you through environment selection, credential configuration, and tool mode setup (minimal or full).

Option 3: Manual Installation

# Follow manual installation guide in docs if you prefer to customize deployment
# This typically involves creating a virtual environment, installing dependencies, and configuring credentials manually

Prerequisites summary: ensure Python is installed, install pip dependencies, and obtain IBM MDM credentials and base URLs as described in the Configuration section of the README.

Additional notes

Notes and tips:

  • Use the uvx approach for Claude Desktop integration to enable STDIO-based communication with minimal setup.
  • Ensure environment variables reflect your deployment (cloud vs. Software Hub) and that MCP_TOOLS_MODE is set to either minimal or full as needed.
  • If you encounter authentication issues, verify API keys, tokens, and base URLs are correct and that the IBM MDM instance is accessible from your network.
  • For debugging, start the server in HTTP mode first to verify endpoints, then switch to STDIO/Claude integration.
  • When using Claude Desktop, you can provide a sample Claude config with mcpServers.ibm-mdm entries benefiting from explicit env variable injections for secure operation.

Related MCP Servers

Sponsor this space

Reach thousands of developers