Game creation is AI-assisted, but the result is still represented as editable game objects, mechanics, and scripts. The goal is to make iteration more understandable and controllable than vibe coding an entire game.

Compared with engines like Unreal, the game maker is intentionally narrow: it focuses on a set of common game mechanics, which makes it much faster to learn and use.

I recommend starting with the demo game. You can play a 5 minute mission immediately, or open it in the editor without creating an account. Both options are available from the front page.

The engine runs in the browser using WebGPU and is written in C++ via Emscripten. All humanoid animations are generated in real time with the help of a physics simulation (Jolt).

Longer term, I’d like to use neural rendering to translate the games’ "semantic pixels" into arbitrary visual styles. Near-term features include music, more varied sound, and a unified agent interface. Currently, behavior creation is split between objects and scripts rather than handled through a single iterative workflow.

There is also a free trial for creating your own game.

Let me know about your thoughts :) It's still early but most fundamental mechanics and features are in place.

Known issue: the current version may crash on iOS 26, fixed on iOS 27. The game can be played on touch/mobile, but the editor designed to be desktop-only.

Thanks for taking a look!