Organization repository on GitHub
jev based AI town simulation
- Forks
- 7
- Last pushed
- Updated Oct 2026
Repo stats from the GitHub API, cached Oct 2026.
CommunityJev Demos & Experiments
JevTown — NevaMind-AI
Pixel-art simulation engine where agents pick structured next actions through Jev while a separate language model handles dialogue, memory, and beliefs, all running in one browser tab.
- Stars
- 46
- Language
- TypeScript
- License
- MIT
- Published by
- Community
- Author
- NevaMind-AI
- Last verified
- Oct 2026
About this project
The first Jev-based AI simulation system. Jev Town is a game engine where the characters are driven by AI, not scripts. Every agent in the world decides what to do next through two kinds of models working together: Jev, a System One model, makes the fast, structured decisions: whether to seek someone out, who to go to, where to wander, how long to wait. The engine offers the options; Jev chooses within them, so an illegal move is never possible.
Quoted from the project's README. Read the full README.
Used inJev for Software & DevTools
- Languages
- TypeScript
Highlights
- Jev makes fast structured moves from offered options so an illegal move is never possible.
- Language model covers conversations, memory, and each agent's beliefs about the world.
- Single-tab simulation with a local proxy holding keys so the browser never sees them.
- Cast size configurable up to dozens of agents, defaulting to a small group for cheap runs.
Watch out
Described as an MVP still taking shape on a feature branch, with every decision and dialogue line costing a model call and a proxy spend backstop that halts after many calls per run. Copy reviewed against the repository README on 2026-10-04, not from running it.
Questions about this entry
Short answers drawn from the linked project or page.
Yes. Every decision and every line of dialogue is a model call, a larger cast costs more, and the local proxy halts after a fixed number of calls per run until restarted.
Yes. Leaving out the action-decider setting hands every decision to the language model instead of Jev, which is useful for comparing the two control modes.
You need Node 22 LTS and npm, a credentials file covering the language endpoint plus Jev key, then a single demo command and the local browser URL.
No. The project calls itself an MVP engine with a game still being built on top, and the Jev demo lives on a feature branch until it merges.
Jev makes the fast structured decisions, such as whether to seek someone out, who to go to, where to wander, and how long to wait. An LLM handles conversations, memory, and what each agent believes about the world.
With Node.js 22 LTS, run npm ci, create .env.local with your LLM values and JEV_API_KEY plus VITE_ACTION_DECIDER=jev, then npm run play:demo and open http://localhost:5173/ai-town/.
The local proxy stops after 2000 model calls per run as a spend backstop; restart it to continue. Every decision and line of dialogue is a model call, so a large cast costs more.
Other listings of this project
Forks and copies of this project in the directory. This entry is the canonical one.
14GitHub stars
A social network where 10,000 computed personas read your post: Jev scores who the text is for to pick the first 600 readers, then answers one Choice per persona for its reaction.
Jev Cookbooks#community#javascript#cloudflare
Community
gaborishka/jevtownJavaScript
gaborishka/jevtownJavaScriptMore Jev resources like this
21GitHub stars
Pi guard plugin that checks inputs and tool calls with local rules plus Jev semantic judgements, with redaction and verification reminders.
Jev Demos & Experiments#community#jevbest#catalog-import
Community
zszz3/Pi-Jev-GuideTypeScript
zszz3/Pi-Jev-GuideTypeScript7GitHub stars
Demo that turns free-text UI descriptions into real shadcn components using only Jev typed Choice, Noul, and Score answers, with no generated code on the decision path.
Jev Demos & Experiments#community#jevbest#catalog-import
4GitHub stars
TypeScript browser harness where an OpenRouter planner splits goals into subgoals, Jev picks each concrete action, OpenRouter fills text fields, and Playwright drives the page.
Jev Demos & Experiments#community#jevbest#catalog-import
Community
vinilana/jev-browserTypeScript
vinilana/jev-browserTypeScript7GitHub stars
Two TypeScript monorepo products that keep the rules in the engine and ask Jev exactly one question: what should this person or this shogi side do next. Hundred runs a hundred NPCs with hunger, money, jobs, relationships and memories; Jev Shogi picks Gote's reply from engine-legal USI moves.
Jev Demos & Experiments#community#jevbest#catalog-import
Community
jammaru/jev-labTypeScript
jammaru/jev-labTypeScript4GitHub stars
Open-source Electron computer-use agent for macOS and Windows that reads the apps you select through native accessibility with local OCR added for text targets, and asks Jev to choose concrete actions that it shows you before executing.
Jev Demos & Experiments#community#jevbest#catalog-import
Community
NobleSpartan6/ottoTypeScript
NobleSpartan6/ottoTypeScript1GitHub stars
A decision board that sends one System One request describing eleven stream ideas, then lets code gate and weight Jev's answers into a single pick for what to build next.
Jev Demos & Experiments#community#jevbest#catalog-import
From the Jev guides
Original write-ups on what this entry is for.
A game tick is a state plus a legal action list. How Clash Royale, Minecraft, and the earlier Mario, Pokémon, StarCraft, and drone loops ask Jev one step at a time.
Jev Demos & Experiments#gaming#agents#control
Read article
From the Jev community
Posts from builders shipping with Jev right now.