{"id":18485,"date":"2026-03-26T13:12:27","date_gmt":"2026-03-26T17:12:27","guid":{"rendered":"https:\/\/blogs.mathworks.com\/simulink\/?p=18485"},"modified":"2026-03-26T14:18:53","modified_gmt":"2026-03-26T18:18:53","slug":"refactoring-a-simulink-model-using-the-matlab-mcp-server","status":"publish","type":"post","link":"https:\/\/blogs.mathworks.com\/simulink\/2026\/03\/26\/refactoring-a-simulink-model-using-the-matlab-mcp-server\/","title":{"rendered":"Refactoring a Simulink Model using the MATLAB MCP Server"},"content":{"rendered":"
\"\"<\/img><\/a><\/div>
Once you start using those AI agents, it's hard to stop!<\/span><\/div>
In today's post, I used <\/span>Amp<\/span><\/a>, an AI coding agent made by <\/span>Sourcegraph<\/span><\/a> and similar to <\/span>Claude Code<\/span><\/a>, to do one of the tasks I help Simulink users with the most: Improve the componentization of a model by using a referenced model in accelerator mode.<\/span><\/div>

The Challenge<\/span><\/h2>
A best practice I always recommend to Simulink users and that we describe in the article <\/span>Improving Simulation Performance in Simulink<\/span><\/a> is to leverage referenced models in accelerator mode. Because they can be precompiled, they initialize instantaneously and simulate faster when they are not modified between simulations.<\/span><\/div>
To highlight the workflow, I started with this example: <\/span>Helicopter Modeling and Simulation - MATLAB & Simulink<\/span><\/a><\/div>
If you want to open this example in your MATLAB, you can execute this command:<\/span><\/div>
openExample(<\/span>'simulink_aerospace\/HelicopterModelingAndSimulationExample'<\/span>)<\/span><\/span><\/div><\/div>
As it is the case in most user models I see, a good candidate to be converted to a referenced model is typically the control logic, in this case the subsystem <\/span>Attitude and Altitude Control<\/span>: <\/span><\/div>
\"\"<\/img><\/div>
While Simulink includes a <\/span>Model Reference Conversion Advisor<\/span><\/a> that can do most of the work, it's common that something does not work as expected after doing the conversion. <\/span><\/div>
I knew that this example had some of those, so I thought it would be a good candidate to test an AI agent.<\/span><\/div>

The Final Result<\/span><\/h2>
Let's first look at a video of how Amp succeeded at the task by iterating through the <\/span>MATLAB MCP server<\/span><\/a>.<\/span><\/div>
What I described is:<\/span><\/div>
  • What to do:<\/span> Convert the <\/span>Attitude and Altitude Control<\/span> subsystem into a referenced model<\/span><\/li>
  • My motivation:<\/span> Improve performance by simulating the referenced model in accelerator mode<\/span><\/li>
  • Requirement: <\/span>I need to be able to change variants in the controller without triggering a rebuild of the accelerator target<\/span><\/li><\/ul>
    I suggested that it should start by creating a baseline test to validate that the final results match the original model and let it iterate.<\/span><\/div>