This is an experimental project exploring how to create a game using AI generation tools, without mastering game engines like Godot or Unity.
The tools involved include AI IDEs Cursor, art tools AI features (Figma, VberAI Studio), and image generation models (GPT Image 2, Nano Banana 2) in t3.chat.
So, here is the breakdown of my AI budget (monthly):
- Cursor: $60
- T3: $8
- Figma: $15 (dev account, just for MCP)
- VberAI: $30
Background:
- Designer, familiar with UI&UX things;
- Developer, familiar with Dev things(basic)
- PM, familiar with product design & plans & AI tools
- Nothing background with Game Design or Dev
I anticipate that this blog will explore the operational limits of VberAI Studio in addressing typical game production scenarios. As I currently understand it, the product leverages an interaction model akin to Figma’s to harness and control game art assets; it then facilitates intensive coordination via game engine MCPs (like Godot MCP from VberAI), thereby decoupling traditional game development tasks from their reliance on human labor.
As I am unfamiliar with many fundamental operations of the game engine, I will document the process with an “AI-first” mindset, attempting to show how an outsider to game development approaches problem-solving. If you encounter any steps that seem baffling, please bear in mind that I am not a professional game developer.
To save time, AI plays a significant role in assisting with the expression of ideas throughout this process. Even if you notice traces of AI-generated writing, please overlook them; this is fundamentally a game development exploration driven by a human, with AI tools simply accelerating the journey of discovery and articulation.
If this journey of exploration offers you any inspiration, please feel free to contact me at liushoapeng@gmail.com.