Setup MTA MCP Server

Purpose

This document describes how to connect your AI assistant to the MTA MCP Server.

The purpose is to design and execute tests using AI, and analyse the test results.

If you specifically want to connect Mendix Maia, checkout the Mendix documentation at: https://docs.mendix.com/refguide/maia-mcp/

Prerequisites

• Make sure that the MTA Plugin module is imported and configured.

Prepare MTA

• Login as Service Account Manager and create a new Service Account.
• Make sure to enable the Call MCP primitive tools scope.
• Add a new Identification token and make sure to copy it into a password manager.

Setup MCP Connection

• Open the configuration or edit the configuration file (for example, mcp_config.json) for the LLM that you are using.
• Enter the URL to the MTA MCP Server, typically https://your-mta-installation.mendixcloud.com/primitivetools/mcp.
• Make sure the URL to the MTA MCP Server ends with /primitivetools/mcp
• Enter the token created above.
• When the connection is succesful, you should see some indication of this, including a list of MCP tools that are enabled.

Loading MCP Skills

Loading the Skill files can be done by importing the Menditect agentic test skills module from https://marketplace.mendix.com/link/component/305457 into your Mendix App, if you are using Mendix Maia. Read more at https://github.com/Menditect/agentic-test-skills#option-2-mendix-studio-pro--maia-users

For other AI assistants (such as Cursor, Claude Code, GitHub Copilot, Cline, or Antigravity / Gemini), clone the Github repository. Follow the instructions here: https://github.com/Menditect/agentic-test-skills#option-1-turnkey-workspace-template-recommended-for-standalone-ai-agents

Feedback?

Missing anything? Let us know!

Last updated 15 September 2026