What it is
A persistent world, not a one-off demo. Industry, economy, and governance live inside the same system and leave connected consequences behind.
Project Overview
oasis7 Project Overview
This is a whitepaper-style explanation for first-time readers. It answers the basic questions: what oasis7 is, why it is worth building, what role the player has, how the world works, and where the current public boundary really sits.
Document type: Explanation / Overview
Public posture: technical preview, not playable yet
This is not a new source of truth
One-page summary
`oasis7` is a civilization simulation game. LLM-driven agents gather, produce, trade, coordinate, and govern inside a continuously running asteroid-belt world. Resources, institutions, and decentralized consensus are not decoration; they are constraints that actually shape what can happen.
Who the player is
The player is closer to an external strategist than a unit-by-unit operator. You set goals, change constraints, and push governance.
What it is not
It is not just a chatbot skin for NPCs, and not a sandbox pushed forward by static scripts. The project is testing whether a world can grow its own institutions, markets, and protocols.
What this page does not replace
This page explains the project, but it does not replace `README.md`, `world-rule.md`, `doc/core/prd.md`, `doc/game/**`, or module-level PRD / project truth.
Why build this
A lot of AI game or agent demos do not fail because the model is weak. More often, the world around it is too soft: agents can talk, but they do not bear consequences; the rules look scripted; the player feels like a puppeteer.
The common gap
There is often no resource conservation, no persistent time, no auditable state, and no durable strategic pressure. The highlight works, but the system underneath does not hold.
The oasis7 approach
Start with hard questions first: does time keep moving, do computation and energy have real cost, where does player influence stop, can rules evolve, and can key state changes be recorded and replayed.
What the project is really testing
If agents have long memory, live under world rules, and can be extended through WASM modules, can a world grow its own organizations, markets, and protocols. Without answering that first, later claims about civilization, governance, or economy tend to collapse back into presentation.
Vision and the two-phase path
The point is not just “more content” or “better tech.” The point is a world that remembers actions, accumulates history, and changes because of what people do inside it.
Participation and achievement
The goal is not a system you only watch. It is a world where people build, test, coordinate, govern, and leave traces that matter later.
The creator role
Creators should not stay in the role of consumers only. If the base becomes stable enough, creators should be able to publish their own `mod / modules` and organize different rule sets and play styles.
Target-state boundary
The longer-term direction is for the game and its key data to move toward something shaped, governed, and benefited from by participants. That is a target-state vision, not a current legal or on-chain fact.
Phase One: open-world sandbox game
First make participation, exploration, and consequence feel strong enough that people want to stay in the same world and keep returning to it.
Phase Two: open-world game platform
Then expand it into an open-world game platform where different rule sets, different mod ecosystems, and different participant roles can grow on the same base layer.
The player, the world, and the game loop
The player is not a world entity. The player is closer to a world-external strategic actor: set high-level goals, adjust prompts and preferences, deploy modules, and push cooperation or governance decisions.
What you can do
Set direction, change constraints, and influence how organizations and systems evolve. The interesting part is the consequence of your judgment, not raw execution speed.
What you cannot do
You cannot directly puppet agents, bypass world rules, or freeze time to optimize every move.
What you should expect
Action-level feedback first, then day-scale shifts in trade and cooperation, then longer changes in institutions and power.
Three qualities of this world
It is alive because time keeps moving. It is hard because movement, computation, and governance choices have cost. It is plastic because many social structures are meant to grow from in-world behavior, not from fixed story rails.
Why the technical base looks like this
The current high-level shape is `LLM Agents -> WASM Modules -> World Runtime -> Consensus Layer -> Distributed Storage & Networking`.
Design intent
LLMs handle decision and interpretation space, WASM carries deployable and auditable higher-level logic, runtime validates rules, and consensus keeps state multi-node verifiable.
Cost and tradeoff
This increases system complexity and slows product closure, but the project would rather move slower than break world consistency for a smoother short-term demo.
Where the project actually is today
It helps to separate three layers: what can be studied, what can be verified, and what can be publicly played. Today you can already understand the direction, inspect the docs and architecture, and run preview validation paths. You should not read that as a public playable release.
Project framing and rule boundaries
The core framing, world rules, player boundaries, and gameplay skeleton already have formal document entry points.
Technical-preview chain
Runtime, viewer, testing, and docs already support build, manual, and validation flows for a technical preview.
Still not a public playable build
Browser verification and downloadable builds do not mean public-scale playability. The current public posture is still technical preview, not playable yet.