Agentic AI for MATLAB and Simulink
Explore the latest tools, examples, releases, blogs, and events
Ebook
Learn how to build AI agents that write, execute, and debug MATLAB code autonomously using the Model Context Protocol, MCP Core Server, and tool calling.
Get ebookIn engineering workflows, agentic AI can help write MATLAB code, run simulations in Simulink, understand results and errors, and iterate. Instead of manually executing each step, you oversee the loop while the agent moves the workflow forward.
Connecting an AI agent to a code interpreter is not unique, but connecting it to MATLAB and Simulink gives the agent engineering-specific capabilities that change the experience:
- Engineering-grade tools: Agents call validated domain-specific toolbox functions for signal processing, controls, and code generation.
- Simulation-driven iteration: Agents test designs against mathematically grounded Simulink models instead of guessing.
- Curated agent skills: Toolkits package engineering expertise so agents write idiomatic code from the first turn.
- Desktop to production: Agents deploy tested MATLAB functions as multi-user tools on MATLAB Production Server. No rewrite necessary.
- No AI-vendor lock-in: Agents work with Claude Code™, GitHub Copilot®, Gemini CLI™, Codex™, Amp™, and others.
Together, these capabilities let agents support everything from writing MATLAB code to running full system simulations in Simulink.
| Tool | Description | Link |
|---|---|---|
| MATLAB Agentic Toolkit | Equips AI coding agents with expert knowledge and a live connection to MATLAB. Provides skills for data analysis, app building, and software development plus over 100 more for specialized engineering domains. | GitHub |
| Simulink Agentic Toolkit | Extends agentic AI into Model-Based Design. Adds Simulink-specific MCP tools and skills, encoding expertise in designing, debugging, and testing. | GitHub |
Gives your agent the knowledge and context to work with Polyspace efficiently. Polyspace MCP Server and companion skills enable AI assistants to perform Polyspace as You Code and Polyspace Test workflows on C/C++ code. |
||
The foundation layer that allows external AI agents (Claude Code, GitHub Copilot, Gemini CLI, Codex, and Amp) to write, execute, test, and analyze MATLAB code in a live session. |
||
Publishes MATLAB functions as MCP tools on MATLAB Production Server for multi-user, enterprise deployment with AES-encrypted IP protection. Clients call tools without needing a MATLAB license. |
||
Enables MATLAB agents to connect to HTTP-based MCP servers, discover available tools, and call them to access external services and extend capabilities. |
||
Connects to your AI agent through a command line interface directly in MATLAB. |
||
| MATLAB AI Agent SDK | Builds AI agents in MATLAB for engineering workflows, connects to externally hosted or local LLMs, exposes MATLAB capabilities as agent tools, and orchestrates agent behavior with full visibility and control. | GitHub |
Blog Posts and Technical Content
Explore the latest technical deep dives and how-to guides on agentic AI with MATLAB and Simulink.
- Using Agentic AI to Design and Deploy a Control System (13:36)
- Video: Can Agentic AI Develop Embedded Systems with Model-Based Design?
- How to Engineer an AI Skill for MATLAB (May, 2026)
- Ebook: Guide to Agentic AI with MATLAB and Simulink
- Automate the Documentation of Your MATLAB Analysis with Agentic AI (Feb 2026)
- Building an Optimization Agent with MATLAB MCP Server (Jan 2026)
- Giving Local AI Agents the Ability to Use MATLAB with MCP (Jan 2026)
Courses
Learn how to work with AI agents in MATLAB and Simulink for engineering workflows.
Applying agentic AI with MATLAB and Simulink across the design lifecycle, from requirements analysis through simulation to deployment.
Workflows
1. Filter Design and Optimization
An AI agent designs, compares, and optimizes digital filters in MATLAB. It proposes filter parameters, evaluates frequency response, and iterates toward specifications automatically.
Using: MATLAB Agentic Toolkit
2. Requirements Engineering and Traceability
Agents parse requirements documents, generate traceable requirements, and verify coverage across functional, logical, and physical architecture layers.
Using: Simulink Agentic Toolkit
3. Thermal Brake Simulation
An agent reads a Simulink thermal model, adjusts brake component parameters, runs simulations, and reports results. Engineers stay in control and let the agent handle the iteration.
Using: Simulink Agentic Toolkit
4. Control System Design and Tuning
AI agents design and tune controllers by interacting with Simulink models, running simulations with different design parameters, and evaluating performance against specifications.
Using: Simulink Agentic Toolkit
