PINNStudio: A free, open-source no-code GUI for setting up, training, and visualizing PINNs [P]
PINNStudio is a free, open-source, no-code GUI designed to simplify the setup, training, and visualization of Physics-Informed Neural Networks (PINNs). It automates code generation using DeepXDE, runs models, streams training logs, and displays live loss curves and solution plots within the application. The tool includes built-in templates for classic equations such as Heat, Allen-Cahn, and Cahn-Hilliard, aiming to reduce the manual effort previously required for modifying PDEs, boundary conditions, and architectures.
- Published
- Sep 6, 2026, 22:19
- Source type
- Dev community
- Tier
- Community
- Source status
- Healthy
Times shown in UTC
More details
When I first started working in scientific machine learning, I understood the physics much better than the coding. Every time I wanted to try a new physics-informed neural network problem, I had to start almost from scratch: changing the PDE, updating boundary conditions, modifying the architecture, tweaking the training schedule, debugging errors, and generating plots—all by hand.
That frustration pushed me to build PINNStudio. It is a free, open-source no-code GUI designed to eliminate boilerplate code so you can focus entirely on the physics.
Instead of rewriting a new script for every problem, you can define your setup directly through the interface:
- PDE Definitions & coupled multi-output PDE systems
- 1D or 2D domains with boundary and initial conditions
- Network architecture & custom training schedules
- Forward problems (solving known PDEs) or Inverse problems (estimating unknown parameters from data)
What happens next? PINNStudio automatically generates the code (built on top of DeepXDE), runs the model, streams the training log, and displays live loss curves and solution plots directly inside the app. It also includes built-in templates for classic equations like Heat, Allen-Cahn, and Cahn-Hilliard.
- GitHub (Open Source): https://github.com/AsfandyarKhan72/PINNStudio
- Quick Install: pip install pinnstudio
My hope is that this will be helpful for students and researchers with limited coding experience, as well as experienced PINN users who just want a faster workflow.
I’d love to get your feedback, feature suggestions, or bug reports! Huge thanks to Lu Lu and the DeepXDE team for creating the foundation that made this possible.