Launches on September 21, 2026

Experimental opt-in Jev skills and config for Claude Code — the companion to jev-codex-tools.

Command-line tool for the TypeSafe Jev decision model, through the official API or OpenRouter: typed yes/no, choice and score questions with calibrated confidence, and confidence gating as exit codes.

High-performance Unix semantic reflex CLI powered by TypeSafe Jev for agent progressive discovery & pipelines

Early evaluation of TypeSafe's Jev on clinic call triage: 98.6% accurate when confident, 116k decisions per dollar

Qualitative coding at scale with Jev: apply a codebook to open-ended text, review uncertain items, measure agreement against your human coders.

Maintain versioned semantic decisions as indexed PostgreSQL columns with a reclaimable work queue.

A demo for Jev (Typesafe AI)

An 8-bit computer built from one yes/no question asked to Jev (TypeSafe) — 24,511 NAND gates from a single API call

TypeSafe Jev (System One) decision layer + CPU-first GUI agent prototype (RapidOCR + Jev + pyautogui + dHash)

Connect Four played by TypeSafe's Jev System One model: one typed Choice per move, with Jev vs Jev personas and a baseline.

Conditional payoff proofs for prediction-market contracts with optional Jev semantic review.

让 Agent 长出「操作杆」：用 System One 模型（Jev）承担高频判断、多模态模型当眼睛。给思路 + 可跑的参考实现（75 条离线断言，零依赖零 key）

An 11-level crash course on Jev, TypeSafe AI's System One decision model — runnable examples against the real API, plus a capstone project with unit tests and evals. Works with any LLM provider via LiteLLM.

Run a declared factorial audience grid through typed Jev reactions and expose disagreement.

Ticket-Routing im Kundenservice mit dem Entscheidungsmodell Jev von typesafe.ai

Resolve ambiguous joins between French public datasets with auditable candidates.

Same maze, same rules, N decision backends. A zero-key, zero-dependency arena that measures Jev / LLM / local decision models on speed, cost, calibration error (ECE) and invalid-output rate.

让 AI Agent 直接调用 TypeSafe AI「Jev」决策模型的 Skill —— 免 API key，纯标准库 CLI。Agent Skill for TypeSafe AI Jev (System One).

A general-purpose decision tool powered by TypeSafe Jev. Users bring their own API key.

Local web app for TypeSafe Jev decision probabilities

A Jev-powered security guard for AI agents: blocks dangerous tool calls, strips prompt injection, scans skills. Works with Claude Code, Codex, Copilot CLI, Gemini CLI, Cursor, and OpenCode.

A concise TypeScript and Bun demo of TypeSafe Jev with live OpenAI comparisons.

a demo of Typesafe's Jev a System One Model for structured decsion making

基于 TypeSafe jev 模型（System One）的桌面端评估工作台

Windows voice desktop agent powered by TypeSafe Jev and Python

Open-source developer harness and runtime safety toolkit for AI coding agents powered by TypeSafe AI / Jev

Researching shared-axis comparison of independent divination readings with Jev / TypeSafe AI.

Classify financial PDFs with TypeSafe Jev via OpenRouter, with a probability per document type.

Run visible criterion-by-criterion text duels and honestly progressive elimination brackets with Jev.

A 2D grid-island ecosystem where each animal is a Jev (TypeSafe) agent; the food chain emerges from calibrated typed decisions. Pure-stdlib Python + browser UI.

A Claude Code plugin that takes skills already in use and adds one thing to each: a typed judgment from [TypeSafe](https://typesafe.ai) Jev (System One) at the points where the skill previously relied on the agent's own impression.

Decide which code-migration sites can be safely auto-rewritten, using Jev's calibrated probability entropy. AST matching, triage, auto-rewrite, self-calibration.

Self-improving agents that know what's noise: typed decisions instead of generated text, a policy that evolves from the agent's own mistakes, and a report of how much of the gain was selection luck.

Experiments with TypeSafe Jev — a Gmail organiser built on typed System One decisions

Trace DeFi incident exposure through sourced portfolio dependencies with optional Jev classification.

Turn documents into typed records with bounded, auditable Jev questions and deterministic aggregation.

Using jev to filter out articles from RSS feed that clog it up.

2D fighting demo where TypeSafe's Jev (System One) makes real-time decisions

Group errors by underlying cause after normalization and exact grouping, with representative comparisons.

Scripts and raw results for a first hands-on look at Jev, TypeSafe AI's System One decision model: calibration, latency, edge cases, and a plain-LLM baseline.

TypeSafe Jev System One powered Gomoku arena

A first-principles rebuild of TypeSafe's Jev: the non-autoregressive System One decision model, dissected and measured against a world whose Bayes posterior is computable

Static code scanner (SAST linter) that finds insecure usage of Jev / TypeSafe guardrails in Python. Not a guardrail itself — build-time analysis, offline, no API key.

A safety guardrail that screens a user message against an assistant's policy before the assistant ever sees it.

Track parliamentary documents that materially affect a declared topic.

Compute probabilities over a heierarchy of terms rather than a set with Jev from Typesafe

Heads-up no-limit Texas hold'em against TypeSafe Jev. The local engine deals and settles; Jev answers typed questions, and code composes the bet.

Jev (TypeSafe) typed prefilter skill: two-phase inbox triage with confidence gating + pre-send guard

A "fun" Godot game that relies on Jev and a local ollama instance to play a text based game on getting out of a Russian Gulag interrogation

Jev CLI for the jev-judge Claude/Codex skill. Wraps the official @typesafe-ai/sdk.