An AI world simulation built from small consequences

A shared world, not six separate conversations

The inhabitants live in the same places and depend on overlapping resources. Mira works at the mill, Bram farms, Nia tends animals, Tomas hauls goods, Sela keeps the inn, and Old Père brings the experience of an elder and former smith. Their roles make another person’s decisions relevant to their own.

The village has a mill, an inn, a market square, a north field, and cottages. Moving between places takes time. Work, food, rest, and encounters therefore happen within a shared setting rather than an unlimited text conversation.

Why the world rules matter

The simulation tracks conditions such as hunger, fatigue, supplies, and ongoing activities. An inhabitant cannot make an unavailable item exist simply by describing it. Actions have prerequisites and effects, and the world validates an attempted action before it becomes an event.

Consider the mill: grain supply, milling work, and the condition of the mill affect one another. A neighbour’s wish for flour does not complete the work. These dependencies create situations in which cooperation, delay, and competing priorities can become visible.

Three layers to look for

• World state: who is where, what resources exist, and which activities are underway.
• Agent decisions: which legal action an inhabitant chooses in the situation, with Jev providing structured judgments.
• Presentation: the three-dimensional scene and narrated chronicle make recorded changes understandable.

Persistence and the pace of village life

The live village runs on a server and saves its history. Closing your browser does not erase the inhabitants or begin a fresh world. The main scheduled decision cycle is ten minutes, with activity and interaction updates between those cycles. The observer follows the latest available state.

Animation and decision-making have different clocks. A figure can walk across the scene without sending a new model request for every frame. If a provider is unavailable or the daily allowance is exhausted, new AI choices may pause; the last recorded state can still be displayed.

What kind of simulation is this?

This is a small social and resource simulation that you observe in a browser. It is not a generator for arbitrary 3D worlds, a physics benchmark, or a forecast of real societies. The inhabitants operate within a finite action vocabulary and a deliberately small environment.

The story can vary because people encounter different pressures and make different choices. That does not mean every possible behaviour is available. The visible world is both an AI experiment and an authored setting, with rules that determine what can happen.

How to explore it

• Choose an inhabitant and follow their public activity across several updates.
• Read the chronicle alongside the scene to connect an action to its outcome.
• Compare neighbours with different roles instead of assuming all agents have the same priorities.
• Open the environment map to understand how the places connect, then return to the live view.