Launches on September 24, 2026

A Pi agent harness built around TypeSafe's Jev (System One model): router, context picker, gate, verifier. Tested on Neon Postgres branches.

Catch an AI agent stuck in a loop — Jev grades each step progress/repeating/stuck/escalate. Single-file, no build. Use free: jevtypesafeai.com/tools/agent-loop-detector

Marks every claim in an arXiv abstract, the sentence that delivers it, and the caveats worth knowing — right on the page. No summaries: every mark is the authors' own sentence. Chrome extension, bring your own key.

Evaluate JEV robot control in MetaWorld and RoboTwin: discrete actions, hierarchical control, privilege-level ablations, and replay analysis.

Find the LLM calls in your codebase that are really decisions, see what they'd save on TypeSafe Jev, and prove it on live traffic before you swap. TypeScript, JavaScript, Python.

Windows Forms playground for Jev System One — visualize and evaluate TypeSafe primitives (Choice, Score, Noul) with a 3-pane editor, preset builders, and simulated API evaluation.

Prototype: AI-assisted NZQA marking from rubric criteria alone, using TypeSafe Jev for guardrails, criterion scores and confidence-based triage. NOT ENDORSED BY NZQA - CONCEPT ONLY

Jev-trading is a desktop application that combines the ultra-fast neural network model Jev (by TypeSafe AI) with an intuitive visual interface. We've created a platform where the power of algorithmic trading is accessible without writing a single line of code, messing with Python scripts, or configuring servers in the terminal.

Independent benchmark of TypeSafe's Jev against GPT-4.1-mini on 200 labelled enquiries. Finding: a calibrated confidence score tells you when to ask a human, not when to ask a bigger model.

Unofficial Jev-powered Zig terminal agent with deterministic policy gates and encrypted audit trails.

an Information-Geometric Analysis Toolkit for any System-one (Jev, Jevlike) agent systems

Is Jev an instrument? Use Jev to finally answer the ultimate question of life: What is an instrument?

Self-contained showcases of JuL (Juste un LLM) — local typed decisions reproducing jevable.com ideas, running on the wemm-4b-4bit model.

Branch on plain-English yes/no questions, batched into TypeSafe System One calls

The asking side of Jev (TypeSafe System One: noul, choice, score) in Rust: client library, mjev CLI, evaluation harness, and Jev reverse engineered from its docs. Talks to Intel-Phi-Jev.

An abstraction for System One models, currently supporting TypeSafe's Jev and Laya. Typed questions, probabilities, and "don't know" instead of a guess, from Python, the command line, or MCP

Find where a folder of Markdown notes contradicts itself, and ask it anything. Every pair of notes crosschecked, every note asked, with Jev via OpenRouter.

Evaluate content for slop and guide revisions with 15 TypeSafe Jev scores. Not an AI-authorship detector.

System 1 vs System 2 AI models, measured: a chain test comparing one-pass Jev with Qwen 3.8 thinking on and off

Chrome-Extension für viele offene Tabs: Schnellsuche, Gruppen, Sortieren, Doppelte, Wächter und Verlauf. Optional mit Jev über OpenRouter oder TypeSafe.

Jev / TypeSafe AI 驱动的 X 信息筛选：按你的规则发现值得读的帖子。开源判断规则库 + 免费 Chrome 扩展。Filter X with your own criteria.

Categorize Actual transactions with Jev

Measured experiments on decision models. Can an open-weight model re-rank search results? Does a token budget explain a benchmark failure? What does one typed decision cost across model families? Every request and response is recorded, and negative results are published as they came out.

Quasar Airlines: prototipo de agentes de comercio (aerolínea + ancillaries + fintech) sobre anthropics/commerce-agents

MCP server and Codex/Claude plugin for TypeSafe Jev typed decisions.

A tree housing every closed question humans or machines ask, answered by Jev (TypeSafe System One): capabilities, defaults, and jaggedness. Not a benchmark.

Open-source Codex skill for prioritizing employers, roles, and sessions at career fairs under limited time.

粘贴聊天，分析互动模式 — 关系场景、十个互动维度与一份关系解读

Typesafe uv project testing the OpenRouter Decisions API (Jev 1.13) against synthetic static-analysis findings

title

macOS disk cleaner using a model cascade: Jev (TypeSafe) classifies hundreds of cache folders for cents, Claude Opus 5 reviews only the uncertain ones. Shows tokens and cost per model. Moves to Trash, never deletes.

AI model routing for Codex subagents. Let Jev choose the model and reasoning effort for each mission, using your own task profiles.

System One typed decisions answered by a separately admitted LLM through TypeSafe's MIT System One adapter.

TypeSafe's Jev System One model as a system-one/decisions provider Cog: typed, calibrated decisions through explicit versioned bindings.

Color picker built on Sanzo Wada's Dictionary of Color Combinations

Jev (TypeSafe System One) sebagai partner keputusan untuk AI agent - agent skill lintas SDK + MCP server + implementasi terukur. Dokumentasi Bahasa Indonesia.

A precision-first RAG orchestrator: filters before it generates, verifies before it answers.

Demos of Jev (System One) by TypeSafe AI: message routing, confidence traffic light, multilingual Tamil

Experimental Dice Chess webhook bot powered by TypeSafe Jev, with adaptive legal-turn selection and no heuristic pre-ranking.

Fail-open two-hop skill suggestion plugin for DeepSeek Harness (DSH): TypeSafe Jev ranks the roster each user turn and injects one-line skill relevance hints

A framework-agnostic Jev decision layer for ESG reporting — reads every claim in a sustainability report against the evidence behind it. Works over GRI, ESRS, BRSR or ISSB. Zero dependencies, 22 tests.

Roteamento automático de incidentes para o time certo usando o modelo Jev (TypeSafe AI) via OpenRouter, com benchmark de 1000 requisições.

🦅 Native macOS Jev proxy and decision observability, in development

Keeps a Plex library under 80% full without deleting what your household still watches. Rust forecaster for *arr.

Honeytrap: a multiplayer social-engineering trap game. Jev (TypeSafe AI) plays the attacker - picking attack-vs-legit, the vector, the target player, and the exact lure, then adapting to each player's weaknesses like Mastermind. 40 scenarios across 8 vectors. Spot the attack or get phished.

Due-diligence engine that uses Jev for probabilistic link routing and Groq for LLM teardowns, cutting token usage by fetching only high-value pages.

A pre-execution interlock for AI agent tool calls: allow, ask, or block before anything runs. Judged by Jev.

Jarvis — assistente de IA pessoal em Google Apps Script: agente ReAct (Gemini), RAG no Drive, voz no Android e julgamentos tipados (TypeSafe/JEV)

Jev decision layer for Hermes Agent

Demo of TypeSafe's Jev decision model: Noul, Choice, Score