execute_matlab_code
Execute MATLAB code and return the results
LABSCO SUMMARY
MATLAB becomes something an agent can call, with the script that produced a result kept when you want it.
Two tools cover the whole loop — describe, generate, run, read the output — and the saveScript and scriptPath pair is what makes it more than a scratchpad, since the code behind a number survives the conversation. What is not here is everything around a session: no workspace inspection, no variable listing, no figure or plot handling, so anything past running code and reading its output sits outside this set.
What it is
A two-tool bridge to a MATLAB installation: one executes code and returns the results, the other generates code from a natural language description.
What you get
- execute_matlab_code runs MATLAB code and returns the results into the conversation
- generate_matlab_code turns a description of what you want into MATLAB code, which execute_matlab_code can then run
- saveScript and scriptPath on both tools, so the code behind a result is left on disk instead of living only in the transcript
Requirements
A MATLAB installation the server can drive, and a writable scriptPath when saveScript is set. Apache-2.0 licensed.
Setup effort
Build from source — git clone https://github.com/WilliamCloudQi/matlab-mcp-server.git
USE CASES
01
Checking a numerical result without leaving the conversation
Run the code against the local MATLAB installation and read the results back, saving the script to a path when the run is one you will want to repeat.
02
Getting a first draft of a script from a description
Describe what the script should do, get MATLAB code back, then run it in the same session and iterate on whatever the output says.
2 tools
Counted once with only startup setup. Other versions, modes or optional credentials may expose a different list.
generate_matlab_code
Generate MATLAB code from a natural language description
BEFORE YOU INSTALL
What to know first
Tools you get
2 tools when we connected.
Requires
- MATLAB installed on your system (MATLAB_PATH points at its executable)
- Node.js v14 or higher
Where it runs
The documented client runs the server on your machine.
CONTEXT COST
Light Standing cost
Rated by the size of the 2 tool definitions that ride along on every turn, not by how many there are
INSTALLATION
Connect your client
Server setup
TERMINAL
git clone https://github.com/WilliamCloudQi/matlab-mcp-server.git
cd matlab-mcp-server
npm install
npm run buildMCP Clients
Claude Desktop
After the build, add this to ~/Library/Application Support/Claude/claude_desktop_config.json on macOS or %APPDATA%/Claude/claude_desktop_config.json on Windows, with the path pointing at build/index.js inside your clone:
CLAUDE DESKTOP · JSON
{
"mcpServers": {
"matlab-server": {
"command": "node",
"args": [
"<YOUR_PATH_TO_MATLAB_MCP_SERVER>/build/index.js"
],
"env": {
"MATLAB_PATH": "<YOUR_MATLAB_PATH>"
},
"disabled": false,
"autoApprove": []
}
}
}Set MATLAB_PATH to your MATLAB executable — usually C:\Program Files\MATLAB\R2023b\bin\matlab.exe on Windows, /Applications/MATLAB_R2023b.app/bin/matlab on macOS, /usr/local/MATLAB/R2023b/bin/matlab on Linux.
