Inspiration
As a long-time lover of the Grand Strategy genre, I’d always yearned to make one but the usual path to making games like those often spans years for large teams. Drawing inspiration from complex medieval strategy titles and the promise of generative AI, Age of Bloodmarch was conceptualized to test the limits of single-developer game creation through end-to-end OpenAI integration.
What it does
Age of Bloodmarch is a medieval grand strategy game built with Unity (URP) that places you in the role of a ruler managing realms, leading armies, forging alliances, and navigating a world brimming with unique events across a massive procedural map. The game is designed to deliver a deep and engaging strategy experience through meticulously optimized procedural systems and intricate mechanics.
How we built it
The entire game was developed and architected by leveraging OpenAI models (GPT-5.5 / Codex) throughout an end-to-end AI development workflow:
- Core Code and Architecture: The entire C# codebase, custom data structures, and state management systems were built and Refactored with the help of GPT-5.5 and Codex.
- Procedural Map Generation: AI was utilized for the design and optimization of complex terrain generation algorithms to produce an seamless map and achieve extremely fast loading times.
- UI and Assets: AI models helped in the design of the User Interface components, concepting of visual assets and organizing the Unity URP rendering pipeline.
- Game Mechanics and Logic: AI decision making mechanics, event chains and balance systems were created by iteratively prompting various models and refining their output.
Challenges we ran into
- Performance Optimization: Loading vast procedural maps and handling a complex structure of C# classes resulted in significant performance issues. With Codex, we refactored some high-cost memory allocations and optimized tight loops, reducing load times drastically.
- Precision and Scale: Carefully maintaining floating-point precision during scale calculations of assets before integer conversion was an intricate UI layout puzzle that was addressed by refactoring the C# code in a way dictated by AI models.
Accomplishments that we're proud of
- The creation of a fully functional, complete grand strategy game prototype solo as Smerss Interactive.
- Successfully building an end-to-end AI workflow, where every aspect from the fundamental code structure to the visual pipeline was developed using OpenAI models.
- Successfully achieving the seamless execution of procedural map generation within the Unity URP framework.
What we learned
- How to utilize Codex and GPT models as full collaborators in repo-level strategic architectural decisions.
- The importance of well-crafted prompts and well-defined reasoning contexts in the process of Refactoring a large body of C# code.
What's next for Age of Bloodmarch
- Steam Release: Final preparations are underway for a commercial launch on Steam.
- Enhanced AI Mechanics: The AI for NPC realms and the event generation chain will be expanded with more advanced AI model workflows.
Extra: If I won, I think I will add ultra graphics, AI civilizations, AI Features etc.
Log in or sign up for Devpost to join the conversation.