questions: { next_action: boolean, danger: score, }
typesafe-mario
A TypeSafe/Jev agent that plays Super Mario Bros. from structured emulator state.

Every project here really calls Jev in its source · refreshed daily
questions: { next_action: boolean, danger: score, }
A TypeSafe/Jev agent that plays Super Mario Bros. from structured emulator state.
questions: { risk: score, }
Camera-only autonomous drone in MuJoCo with a small judgment model (TypeSafe Jev) in the loop at 2.5Hz
questions: { team: choice, human_requested: boolean, refund_requested: boolean, // +3 }
Interactive experiments with TypeSafe Jev, from support routing to 3D driving simulations with real AI decisions and visible sensor inputs.
questions: { lane_action: choice, speed_action: choice, hazard: score, // +1 }
2D autonomous car simulation in the browser, driven by TypeSafe's Jev decision model
questions: { throttle: choice, yaw: choice, pitch: choice, // +3 }
This demo uses Jev from TypeSafe AI to autonomously fly a drone in a random city from point A to point B, avoiding obstacles along the way. A trip costs $0.01.
questions: { line: choice, pace: choice, overtake: boolean, // +1 }
Uses jev to control race cars on a virtual track
questions: { next_move: choice, }
Chakravyuha — a polar ring-maze where every move is a Jev (TypeSafe System One) decision. A fun experiment: the model picks each move, the walk grades it green or red, and the history page asks whether its confidence score can be trusted. BYOK, no build step.
questions: { intent: choice, connection: choice, supported: boolean, }
A small detective escape room built with TypeSafe Jev, React, and Express.
"jev (TypeSafe stepper)": ("typesafe_sdk", "quackd[jev]"),One CLI for all your robots. Connect them, command them, and let them work together, each with an LLM for a brain, Jev for cheaper steps. Microduck, Open Duck Mini, LeRobot, XLeRobot, AlohaMini, ToddlerBot or any ROS base. Claude, OpenAI, Gemini, Grok, or local via Ollama or vLLM. Simulator, .duck safety contracts, MCP, memory between runs, flocks.
const res = await fetch("https://api.typesafe.ai/v1/systemone", {A playable Three.js driving simulator with Jev-powered autopilot
import { TypeSafeClient, choice } from "@typesafe-ai/sdk";Snake auto-played by TypeSafe's Jev model: one System One choice per tick, legal moves and facts generated in code
API_URL = "https://api.typesafe.ai/v1/systemone"
TypeSafe Jev controls original StarCraft shareware through keyboard and mouse with recorded action probabilities.
"typesafe": ("https://api.typesafe.ai/v1/systemone", "jev-latest", "TYPESAFE"),Fine-grained robot control with Jev, physics previews, and configurable LIBERO tasks.
TypeSafeClient,
Observable browser stealth game: Jev makes typed guard judgments while deterministic code owns the world.
import { choice, noul, TypeSafeClient, type EntryType } from "@typesafe-ai/sdk";1v1 Jev quickscope arena — Three.js + TypeSafe System One
export const TYPESAFE_BASE = process.env.TYPESAFE_API_BASE || "https://api.typesafe.ai";
Jev play Tetris in real-time against other AI models
try{const response=await fetch('https://api.typesafe.ai/v1/systemone',{method:'POST',signal:controller.signal,headers:{'content-type':'application/json','authorA browser-native Doom agent experiment with structured spatial state, composable AI controls, live decision telemetry, and a Chocolate Doom WebAssembly runtime.
reply = requests.post("https://api.typesafe.ai/v1/systemone", json=request,Jev plays Generation 3 Pokémon via Showdown and a real FireRed ROM.
const res = await fetch("https://api.typesafe.ai/v1/systemone", {22 Jev models, one ball: a 3D football match where every player is its own Jev (TypeSafe AI System One) decision. Watch, or take over the number 9.
from typesafe_sdk import Choice, Noul, RetryPolicy, TypeSafeClient
Playing Pokemon Red using TypeSafe Jev
from typesafe_sdk import AsyncTypeSafeClient, Choice, RetryPolicy
TypeSafe's Jev model plays Vampire Survivors on Steam: BepInEx plugin + Python brain + live decision dashboard. Native Linux only.
endpoint = "https://api.typesafe.ai/v1/systemone"
Zero-shot English goals on a sim Franka. Jev chains hardcoded primitives.
export const MODEL = 'typesafe/jev-1.13';
Chess, Connect Four, and a decision model. Play Jev or watch Jev play itself.
import { choice, noul } from '@typesafe-ai/sdk';FUn little experiment with Typesafe AI Jev Model playing chess against stockfish :)
API_URL = "https://api.typesafe.ai/v1/systemone"
TypeSafe'in Jev karar modeli gerçek bir online 2048 sitesinde oynuyor — hamle başına tek API çağrısı, tek anahtar.
return {"url": "https://api.typesafe.ai/v1/systemone", "key": os.getenv("TYPESAFE_API_KEY", ""), "model": os.getenv("TYPESAFE_MODEL", "jev-latest")}EmbodiedJev: MuJoCo robot decision workbench with MiniCPM5-2B, Jev and compatible model APIs
typesafe:Object.freeze({id:'typesafe',label:'TypeSafe',endpoint:'https://api.typesafe.ai/v1/systemone',modelsEndpoint:'https://api.typesafe.ai/v1/models',envKeyAI 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.
import { choice, TypeSafeClient } from '@typesafe-ai/sdk';100 AI NPCs live in a tiny town. Jev chooses the next action; the world writes the story.
API_HOST = "api.typesafe.ai"
A maleCNS fly-brain connectome fights a language model in mk.js — spiking simulation, dopamine learning, and the controls that say what each side contributes
import { choice, type EntryType, TypeSafeClient } from "@typesafe-ai/sdk";JevArena — two Jev agents duel in click-only browser games (Browser Use + TypeSafe Jev)
DEFAULT_BASE_URL = "https://api.typesafe.ai"
Pokemon Red on PyBoy: code owns the route and the arithmetic, Jev picks at branches in about 100 ms, calibration measured instead of assumed
JEV_URL = "https://api.typesafe.ai/v1/systemone"
TypeSafe Jev plays Super Mario Bros from a text description of emulator RAM
* POST /v1/systemone Jev at TypeSafe, which refuses requests from browser pages
Jev, TypeSafe's System One model, plays chess against any OpenRouter LLM, Stockfish and you. One-page web app with live moves, Jev's move probabilities, saved games and win rates.
from typesafe_sdk import Choice, Noul, TypeSafeClient
Jev (TypeSafe System One) vs Claude Opus 5 driving a simulated robot arm in MuJoCo
if event != 'jev':
An experiment in fast, probabilistic decisions. Jev chooses actions; Python handles
`,e+=" a b c d e f g h",e}perft(e){let t=this._moves({legal:!1}),n=0,i=this._turn;for(let r=0,a=t.length;r<a;r++)this._makeMove(t[r]),this._isKingAttChess moves, evaluations, persona opponents, and game classification with TypeSafe AI System One
response = client.post('https://api.typesafe.ai/v1/systemone', json=payload, timeout=min(20, remaining))Experimental Jev game agent: RAM, emulator lookahead, checkpoint search and verified recordings. Bring your own ROM and BizHawk.
const API_URL = "https://api.typesafe.ai/v1/systemone";
Jev plays browser table tennis in real time: structured telemetry, typed decisions, ordinary Chrome inputs, and auditable evidence.
const TYPESAFE_URL = 'https://api.typesafe.ai/v1/systemone';
Jev learns to play flappy-bird game with physics based context and without it
Endpoint: POST https://api.typesafe.ai/v1/systemone
Let Jev (TypeSafeAI) solve 2048
import type { JsonValue, Questions } from "@typesafe-ai/sdk";A living-world RPG whose NPCs are decided by TypeSafe's Jev judge model; code owns rules, numbers and state.
import type { SystemOneResult } from "@typesafe-ai/sdk";Jev (TypeSafe AI) PoC through Game of Thrones
API_URL = "https://api.typesafe.ai/v1/systemone"
TypeSafe Jev plays original Civilization II in a browser, with live action probabilities. Experimental full-game harness.
const API_URL = "https://api.typesafe.ai/v1/systemone";
TypeSafe Jev arcade
endpoint: str = "https://api.typesafe.ai/v1/systemone"
用 TypeSafe Jev(System One)驱动扫雷自动通关:HTML 扫雷页每走一步实时问模型,并把它的选项、概率、代码已证明的事实全部可视化。Play Minesweeper with TypeSafe Jev (System One): an HTML game that asks the model every move and visualises its options, probabilities and the facts the code proved.
import { choice, TypeSafeClient } from "@typesafe-ai/sdk";8 AI snakes, 1 human, 1 arena. Every snake is driven live by TypeSafe's Jev, making all decisions in real time
const UPSTREAM_URL = 'https://api.typesafe.ai/v1/systemone';
トランプの スピード を、CPU 側の頭脳に TypeSafe AI の Jev を使って対戦する WebGL デモアプリ。Jev の判断速度と判断精度をリアルタイムに計測して見せることが目的。
def __init__(self, model="jev-latest", url="https://api.typesafe.ai/v1/systemone", threads=4):
화재 피난자 개인별 판단 엔진 — 타입 안전 판독(Bonsai 2 27B / TypeSafe Jev) + 소셜포스 이동, FDS-GPU 화재장 위에서 역할·구조·경로 결정 · Meteor Simulation
const response=await fetch('https://api.typesafe.ai/v1/systemone',{method:'POST',headers:{Authorization:`Bearer ${process.env.TYPESAFE_API_KEY}`,'Content-Type':弈瞬:双 Jev 五子棋九宫格输入实验台,逐手查看模型决策,支持真实对局回放与实时对战。