For the past few days, I’ve been building something that I believe can quickly become one of the favorite features in the entire Mathify ecosystem. It started as a small experiment, but it immediately opened a much bigger door:
What if Manim scenes could move beyond video—and become fully interactive scientific visualizations?

Turns out, they can.

Manim (the Mathematical Animation Engine) is traditionally a video generator. You write a script, you render a scene, and you get a beautiful animation. But underneath the visuals, there is structure: points, surfaces, parametric shapes, meshes—data.

Once you think of Manim objects as data, something interesting happens:

You can export them.
You can inspect them.
You can interact with them.

That idea is what led me to build ManimVTK — a fork and extension of the Manim Community Edition that adds a VTK export pipeline. VTK (Visualization Toolkit) is one of the strongest standards in scientific computing, used for everything from fluid simulations to medical imaging.

With this extension, the same objects you animate in Manim can now be explored interactively in the browser.

A surface that used to be a flat video?
Now you can rotate it, zoom into it, examine curvature, or—soon—adjust parameters live.

A few users asked for parameter-controlled visualizations: sliders for variables, live manipulation, exploring geometry as if it were a scientific tool rather than a static film.

That resonated immediately, because Mathify’s long-term vision is not just “AI makes a video.”
It’s AI assists you in deep insight by letting you explore mathematical structures.

VTK is, admittedly, a large hammer for this problem—but it’s the right kind of hammer:

ManimVTK is simply Manim + VTK export capabilities, a superset. Same syntax, same workflow—just more power.

This was also the first time I leaned heavily on an AI agent across a large codebase.
Refactors, tests, structural changes, debugging the VTK pipeline—all orchestrated through GitHub Copilot.

It didn't replace me, but it compressed the iteration cycle dramatically.
What would have taken days of mechanical changes happened in hours.

It felt like being the tech lead of a team of engineers I could direct and coordinate.

Even at this early stage, the feeling of taking a once-static animation and suddenly touching it in 3D is surprisingly powerful.

Soon, animations will run directly inside the interactive viewer as well—you’ll be able to pause a scene, interact with its objects as if they were tangible, and adjust parameters through built-in UI controls.

ManimVTK is now published on PyPI

The repository is public, and I’ll be adding documentation, tutorials, and deeper explanations of the design choices.

If you're a Manim user, a scientific-computing person, an educator, or someone who just likes mathematical toys and artifacts—this is something you'll probably enjoy experimenting with.

A deeper breakdown of:

Coming soon.

If you have ideas or want to see certain features, I’m always listening.

Thanks for reading —
José
Founder, Mathify.dev