
One prompt builds the world. One response runs the simulation. No compile step. No asset pipeline. Just language in, game out.
Try the ArcadePrairie Engine turns a language model into a stateful simulation engine. You describe a world in natural language — its rules, its entities, its physics — and the engine maintains that world across turns. Every player input mutates state. Every response is a rendered frame.
There is no game loop in the traditional sense. The model holds state in context, applies rules from the system prompt, and produces output that is both narrative and mechanical. The prompt is the source code. The conversation is the executable.
Starfall Legion is a fleet command game. Sovereign is a grand strategy game inspired by Civilization. Both run on Prairie Engine — same runtime, different context layers. The engine handles turn resolution, state tracking, combat calculations, resource management, and narrative generation in a single pass.
Describe a game that doesn't exist yet. The engine will build it and run it in the same conversation. Genre doesn't matter. Complexity scales with the prompt.
The engine is not limited to games. Anything that can be modeled as state + rules + input → output is a valid simulation. Training scenarios for medical staff. Financial market models. Logistics planning. Urban development. Negotiation practice. Courtroom simulations.
If you can describe the system, Prairie Engine can run it. The model becomes the physics engine, the renderer, and the game master simultaneously.
Prairie Engine separates the runtime from the content. The engine itself is a fixed system prompt that handles state management, turn processing, and output formatting. The context layer is the interchangeable part — a document that defines the world, its rules, and its vocabulary.
Swap the context layer and the engine runs a completely different simulation. Same infrastructure, different world. This is how Starfall Legion and Sovereign share an engine but feel nothing alike.
For organizations, Prairie Engine is a simulation platform. Write a context layer that models your domain — your processes, your decision trees, your edge cases — and deploy it as an interactive training environment, a planning tool, or a decision support system.
Patent pending.