Now in Beta — Free to Use

Optimal control,
intelligently solved.

GPOPS AI translates your problem description into complete, runnable GPOPS-II MATLAB code — in seconds. Built for aerospace engineers and researchers.

Try GPOPS AI Free → See how it works
GPOPS AI — Hydra 1.0
YOU
Set up a minimum-fuel lunar soft landing with a 10 km initial altitude and touchdown at zero velocity.
% moonLanderMain.m — Minimum Fuel Soft Landing clear all; close all; clc auxdata.g = 1.5; bounds.phase.initialstate.lower = [10000, -2]; bounds.phase.initialstate.upper = [10000, -2]; bounds.phase.finalstate.lower = [0, 0]; bounds.phase.finalstate.upper = [0, 0]; bounds.phase.control.lower = [0]; bounds.phase.control.upper = [3]; % ... complete setup, continuous, endpoint files generated output = gpops2(setup);
Beta is free and open. No account required. No credit card. Just open GPOPS AI and describe your first problem.
Capabilities
Everything you need to solve optimal control problems

From simple single-phase problems to complex multi-phase trajectories — describe it in plain English, get runnable MATLAB.

Natural Language Input
Describe your optimal control problem in plain English. No structured templates, no special syntax — just tell GPOPS AI what you need.
3-File MATLAB Output
Generates all three required GPOPS-II files — main, continuous, and endpoint — validated and ready to run with gpops2(setup).
Multi-Phase Support
Handles staging events, reentry phases, and orbital transfers where dynamics change at interior points — all in a single conversation.
20+ Benchmark Problems
Built-in validated examples spanning lunar landing, reentry, orbit transfer, robotics, and more — use them to test or as starting points.
Iterative Debugging
Paste error output back into the conversation. GPOPS AI maintains full context and adjusts bounds, mesh, or structure to fix convergence issues.
Powered by Claude Sonnet
Built on Anthropic's Claude Sonnet with GPOPS-II documentation and validated examples baked into every response.
How It Works
From problem description to runnable code

Four steps. No GPOPS-II expertise required to get started.

01
Describe your problem
Tell GPOPS AI your dynamics, objectives, and constraints in plain English. No templates or structured input required.
02
AI generates all 3 files
GPOPS AI produces main, continuous, and endpoint .m files with validated bounds, consistent initial guesses, and mesh configured.
03
Export and run
Download the .m file directly or copy to clipboard. Open in MATLAB and run gpops2(setup) — no reformatting needed.
04
Iterate in session
Ask follow-up questions, request mesh adjustments, or paste error output back in to debug — full context maintained throughout.
About

Built on decades of optimal control research

GPOPS AI is a natural language interface for GPOPS-II — the General Pseudospectral Optimal Control Software developed by Dr. Anil V. Rao at the University of Florida.

GPOPS-II is the most capable and rigorously validated optimal control solver available, used by NASA, Lockheed Martin, Boeing, and research universities worldwide. It transcribes continuous optimal control problems into large sparse NLP problems using the Radau pseudospectral method.

GPOPS AI makes that power accessible through natural language — reducing setup time from days to minutes for engineers who need answers, not hours of debugging.

Try it free →
20+
Built-in benchmark problems
3
MATLAB files per response
∞
Queries during beta
0
Account required to start

Free during beta.
Try it right now.

No account required. No credit card. No setup. Just open GPOPS AI and describe your first optimal control problem.

Launch GPOPS AI → Contact us