Matlabmcp

This project implements a MATLAB Model Context Protocol (MCP) server, supporting interaction with a shared MATLAB session through a Python interface. It provides functions such as code execution and variable retrieval, and plans to expand support for more tools and data types.

Artificial intelligence chatbots Developer tools #MATLAB Integration #Asynchronous Communication #Scientific Computing #Tool Expansion .Python

rating : 2.5 points

downloads : 5.3K

update time : 2025-04-28

Open Site

What is MATLAB MCP Integration?

This service acts as a communication bridge between AI assistants (like Claude) and MATLAB. It allows you to execute MATLAB commands, retrieve variables, and interact with MATLAB's computational environment directly through natural language prompts.

How to use this service?

After setup, your AI assistant can automatically send MATLAB commands through this server. You simply ask the assistant to perform MATLAB operations, and it handles the technical communication behind the scenes.

Ideal Use Cases

Perfect for data scientists and engineers who want to combine AI assistance with MATLAB's powerful computational capabilities. Use it for quick calculations, prototyping algorithms, or analyzing workspace data without leaving your chat interface.

Key Features

MATLAB Code Execution

Run any MATLAB command or script snippet directly through your AI assistant

Workspace Data Access

Get values of variables from MATLAB's workspace in structured format

Non-Blocking Operations

Commands run asynchronously so your chat interface remains responsive

Shared MATLAB Session

Connects to your existing MATLAB instance, preserving your workspace and settings

Advantages

Seamless integration between AI assistants and MATLAB

No need to manually copy-paste code between applications

Preserves your existing MATLAB workspace and environment

Structured results that are easy for AI to process and explain

Limitations

Requires MATLAB to be installed and running separately

Currently limited in handling complex MATLAB data types

Doesn't support Simulink interaction yet

Requires initial one-time setup configuration

Setup and Usage Guide

Install MATLAB Engine

Ensure MATLAB R2023a or later is installed with the Engine API for Python

Set Up Python Environment

Create and activate a Python virtual environment with required dependencies

Configure Claude Desktop

Add the server configuration to Claude's MCP settings file

Start Using

Restart Claude and begin sending MATLAB commands through natural language

Practical Examples

Matrix Calculation

Compute eigenvalues of a matrix

Data Visualization

Create a simple plot

Workspace Inspection

Check variable values

Frequently Asked Questions

Why can't the server connect to MATLAB?

Can I use this with MATLAB Online?

How do I see the server logs?

What MATLAB data types are supported?

Additional Resources

MATLAB Engine API Documentation

Official documentation for MATLAB-Python integration

GitHub Repository

Source code and issue tracking

MCP Protocol Overview

Technical details about Model Context Protocol

🚀 MATLAB MCP Integration

This is a Model Context Protocol (MCP) server implemented for MATLAB. It enables MCP clients (such as LLM agents or Claude Desktop) to interact with a shared MATLAB session via the MATLAB Engine API for Python.

🚀 Quick Start

This MATLAB MCP server allows MCP clients to interact with a shared MATLAB session through the MATLAB Engine API for Python. Here's a brief introduction to its functions and how to get started.

✨ Features

Execute MATLAB Code: Use the runMatlabCode tool to run arbitrary MATLAB code snippets.
Retrieve Variables: Use the getVariable tool to obtain variable values from the MATLAB workspace.
Structured Communication: The tools return results and errors as structured JSON, facilitating programming handling on the client side.
Non-blocking Execution: Use asyncio.to_thread to asynchronously run MATLAB engine calls, preventing server blockage.
Standard Logging: Utilize Python's standard logging module, outputting to stderr for visibility in the client logs.
Shared Session: Connect to an existing shared MATLAB session.

To-Do List:

Add the setVariable tool to write data to the MATLAB workspace.
Add the runScript tool to directly execute .m files.
Increase workspace management tools (e.g., clearWorkspace, getWorkspaceVariables).
Expand the matlab_to_python helper function to support more complex data types (structures, cell arrays, objects).
Add support for interacting with Simulink models.

📦 Installation

Clone this repository:

git clone https://github.com/jigarbhoye04/MatlabMCP.git
cd MatlabMCP

Set up a Python virtual environment (recommended):

# Install uv if not already installed: https://github.com/astral-sh/uv
uv init
uv venv
source .venv/bin/activate  # On Windows, use: .venv\Scripts\activate

Install dependencies:
```
uv pip sync
```
Ensure MATLAB is installed and the MATLAB Engine API for Python is configured. Refer to MATLAB Documentation.
Start MATLAB and share its engine: Run the following command in the MATLAB command window:
```
matlab.engine.shareEngine
```
You can verify sharing by running matlab.engine.isEngineShared in MATLAB (it should return true).

📚 Documentation

Configuration

Instructions for integrating this server with Claude Desktop:

Clone this repository in your project.
Fill in the necessary configuration information, such as API keys and endpoints, in the config.json file.
Start the server and connect to the Claude platform.

Development

For detailed documentation on the server's functionality, usage, and development considerations, refer to [here](Link to documentation).

🤝 Contribution

Contributions are welcome! If you have any suggestions or improvements, feel free to submit issues or pull requests. Let's make this project better together!