ORIGIN-CIVILIZATION
AI life-and-civilization simulation where Jev makes every person's next action as an auditable Choice and LLMs only plan.
Overview
ORIGIN / CIVILIZATION is an AI-driven life and civilisation simulation in which every voluntary action is a typed decision by TypeSafe Jev. Each person's next action is a Choice over the actions actually available, returning a probability for every option and a confidence. Optional LLM planners, via any OpenAI-compatible API, local models such as Ollama or LM Studio, Claude Code, Codex or agy, propose goals, intentions and speech grounded in memories, but never decide.
Code validates every answer before anything happens: it must identify a Jev model, select an existing option, give probabilities summing to 1 and choose the highest-probability option. Every call is stored as a ledger receipt. There is no local-policy fallback or synthetic response: if Jev cannot answer, the world pauses. Version 6.1 uses HTML, JavaScript and an authoritative Node.js server with an isometric renderer and optional Three.js view. It is original work, not an EA product.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Choice | What should this person do next? | — |
How it works
- Compute the feasible actions for a person.
- Optionally let an LLM planner propose goals.
- Ask Jev a Choice over feasible actions.
- Validate the answer, record a receipt and execute.
Strengths
- Auditable ledger of every decision.
- Strict separation of planning and deciding.
- No fake fallbacks.
Caveats
- Requires a TypeSafe key; the world pauses without Jev.
- Simulation, not a reusable library.
At a glance
- Type
- Project
- Status
- prototype
- Language
- JavaScript
- Runtime
- Node.js server, browser
- Has code
- yes
- Has tests
- yes
Use cases and integrations
How curated lists describe it
AI life-and-civilization simulation: TypeSafe Jev makes every decision (typed, probabilistic, auditable); LLMs plan — OpenAI-compatible APIs, local models (Ollama, LM Studio), Claude Code, Codex.
hellogumbo/awesome-jev
Frequently asked questions
What does Jev decide?
Every person's next voluntary action from the feasible options.
What do LLMs do?
They can plan goals and speech but never decide actions.
What happens if Jev is unavailable?
The simulation pauses; there is no fallback policy.
Related projects
Doom runs as a spacecraft payload over real flight software; Jev pilots from the ground every half second while Claude reviews.
ProjectprototypePythonFactorio agent on FLE where Jev picks goals, next actions and stuck detection while code owns rules and actuation.
ProjectresearchPythonReal-time bot battles where you write a fighter's strategy in plain English and a System One model pilots it several times a second.
ProjectbetaTypeScriptBot that plays Red Alert 2 web skirmishes, with code handling mechanics and a decision model making stance, threat and timing calls.
ServicebetaPythonMultiplayer 3D science-fantasy version of Houston's River Oaks District with Jev-powered NPC companion, chauffeur and smart drive.
ProjectbetaTypeScriptBrowser 3D planetary rover sandbox where a TypeSafe controller chooses the next action from complete candidates, versus a baseline.
ProjectbetaTypeScript