Premise: TestComplete MCP lets your AI assistant explore your project and run tests in plain language. The server runs locally, so your data stays under your control.
QA teams in banking, healthcare, government, and manufacturing are hearing the same thing from leadership: use AI. They are also hearing the same thing from security: do not send application data to the cloud.
TestComplete MCP resolves that tension. It connects the AI assistant your team already likes to TestComplete through a server that runs on your own machine. You get the speed of AI. You keep control of your data.
What is TestComplete MCP?
The Model Context Protocol (MCP) is an open standard that lets AI assistants work with external tools. TestComplete now includes an MCP server, so you can ask an AI assistant in plain language to inspect your test project, run a test, or explain what you have, without leaving the AI tool you are working in.
It ships inside the standard TestComplete installer. There is no separate download and no extra plugin to install. Turn it on in TestComplete, copy a connection URL into your AI client, and you are connected in minutes.
Your data stays in your control
For teams testing sensitive applications, where the server runs matters as much as what it does.
The TestComplete MCP server runs locally, on the same machine as TestComplete. It is not a cloud service. There is no hosted layer between your AI client and your test project.
That gives you control in three ways:
-
You decide when it runs. MCP is an optional feature, and it only runs while TestComplete is open.
-
You choose the AI client. Use any MCP-compatible client that fits your organization’s policies. What you ask the assistant to read is shared only with the client you picked, which makes that choice yours to make.
This fits how regulated teams already like to work with TestComplete: on your network, on your terms, and easier to take through a security review.
What you can do with it today
Once connected, you ask in plain language. Here is what the assistant can do now, and why it matters.
| Ask your AI assistant to | Why it matters |
|---|---|
| List your keyword tests and script tests | Understand a project in seconds instead of clicking through it. |
| Show the code for a script test | Read and explain scripts you did not write, and get unstuck faster. |
| Show the execution plan | See what runs, and in what order, without opening the plan. |
| List name mappings | Understand which UI objects your tests rely on. |
| List project and test variables | Review configuration at a glance. Secrets stay hidden. |
| List tested applications | Confirm what the project is set up to test. |
| Open or close a project suite | Point the assistant at the right project without switching windows. |
| Run a script test or keyword test | Start a run from the same conversation. |
After a run, ask for the test logs with timestamps and result status, or a step-by-step breakdown of what happened. You see what passed and what failed without leaving the conversation.
TestComplete takes control of the real mouse and keyboard when a test runs. Your assistant will warn you before it starts one, so you can step away.
Who benefits
Automation engineers spend less time navigating and more time solving problems. Ask what a script does, check which variables a test uses, and start a run without leaving your AI tool.
QA leads get visibility in plain language. Ask what exists in the project and how it is organized, with no tour of the UI required.
Manual testers and new TestComplete users get a patient guide. Instead of learning every corner of the product first, they can ask what is in the project and what to look at next. For hybrid teams that mix manual testers and automation engineers, this lowers the barrier to getting started with automation.
And because TestComplete supports both scripted and scriptless testing for desktop and web applications, the whole team works in one tool, with one AI assistant on top of it.
What is coming next
This release is about exploring your project, running tests, and reviewing results. We want to be clear about what it does not do yet.
- Test creation and test maintenance. The MCP server cannot create or maintain tests yet. These capabilities are planned for the next couple of months. When they arrive, the same conversation that explores your project will help you build and maintain tests, under the same local model.
Starting now means your team has the connection in place, and the habits built, when those capabilities land.
Get started in four steps
-
Make sure your TestComplete license includes the IQ add-on that MCP requires.
-
In TestComplete, open the MCP settings and switch MCP on.
-
Copy the connection URL that TestComplete displays.
-
Add it to your AI client, then ask: “Can you check if you are connected to TestComplete?”
Then ask something real: “List the keyword tests in this project.”
The bottom line
AI in testing does not have to mean giving up control. TestComplete MCP runs locally, protects your secrets, and lets your team work faster in plain language today, with test creation and maintenance on the way.
TestCompletes MCP quick demo: TC_MCP_Quick_Info