2026년 9월 23일 출시

ollaya는 이름으로 공개 결정 모델을 내려받고 로컬 데몬을 통해 타입이 지정된 질문을 처리합니다. 기존 Jev 클라이언트가 쓸 수 있는 TypeSafe 호환 엔드포인트를 구현한 독립 프로젝트입니다. CLI, 데스크톱 앱, MCP 서버를 이용할 수 있으며 Modelfile로 재사용할 질문 묶음을 정의합니다.

Valen은 텍스트, 이미지, 영상에 대해 주어진 후보를 평가하고 답변 토큰 없이 판단 확률을 반환합니다. Jev에서 영감을 받은 구현으로 Qwen3.5 기반 모델과 공유 결정 헤드를 사용합니다. 추론이나 평가에는 Valen 체크포인트와 Qwen3.5-2B 기반 모델을 모두 준비해야 합니다.

Jev Browser — indexed browser automation. Jev chooses the control, Playwright acts. A CodexQA skill.

A self-hosted API and web interface for Laya’s structured decision models, compatible with the TypeSafe Jev API format.

A local, API-compatible replacement for Jev, TypeSafe's System One model

A database-agnostic Agentic GraphRAG framework using swappable System One models (local Laya / cloud Jev). A plug-and-play intelligence layer featuring a complete 4-phase pipeline, continuous evaluation and custom A* traversal for any graph database.

Go SDK and CLI for TypeSafe Jev: typed decisions (yes/no, choice, score) from a model

Local proxy that learns your app's typed LLM decisions and answers them with a Laya head. Jev and OpenAI compatible.

MoJev: typed, calibrated decisions in one forward pass.

Typed judgment tools for MCP agents. TypeSafe's Jev model as verify, screen, find, classify, rerank, decide, compare, extract, review, gate, and score: the model judges, policy decides auto, review, or escalate.

Research design for a Jev-native agent system:enable more options than jev provided with virtulization and paging, tool integration, external helper logits Top-k proposals with Jev-controlled fallback ,decision-aware hierarchical memory, and dependency-aware replanning.

Let Jev read every new arXiv paper each morning and surface the few you should read. Plain-English interests, calibrated probabilities, ~$0.06/day, fork and go.

Typed decisions from unstructured state: one forward pass, zero generated text. Local, deterministic, Jev-compatible. Not affiliated with typesafe.ai.

iPhone 键盘版：一个自定义键盘打通所有聊天 App——长按复制对方消息，键盘上出意图、风险与候选回复，点一下进输入框。只读剪贴板，源码形式（需自行用 Xcode 编译安装）。

Voice control for macOS that acts partway through your sentence, built on Jev (TypeSafe)

Routes each request to the right AI agent or skill before your AI picks one. Your rules decide what code can; Jev (TypeSafe) makes the typed call. Installs as Claude Code hooks.

弦外之音 —— 微信聊天里的关系进展助手：读屏（无障碍树 / 截屏视觉）→ Jev 判读 + 攻略度 → 聊天模型出 3 条候选回复，攻略度常驻挂在屏幕上。不改微信、不发消息、不注入点击。

Jev-shaped (TypeSafe System One) classification wrapper over OpenAI-like clients

Learn Jev end to end: a free hands-on course. Build 13 AI agent use cases with a fast brain (Jev) and a slow brain (LLM). One OpenRouter key.

Prompt Claude Code: tre sistemi costruiti su Jev di TypeSafe — routing del modello, compattazione del contesto e code review

Let Jev steer Claude Code: the right reasoning effort, subagent model and skill for every prompt. A Claude Code plugin powered by TypeSafe's Jev (OpenRouter / TypeSafe).

Skills without the context tax. Claude Code and Codex plugin: TypeSafe's Jev decides on every turn which skills the model sees. Always-on context cost: 0 tokens.

When and Where Actually to use Jev in your code base.

Run Laya locally with Bun: native ONNX inference, a TypeSafe-compatible API, and a bilingual decision playground.

Jev agent authorization for MCP tool calls: Kinde identity and permissions plus Jev's typed, calibrated decisions, checked server-side before every call runs

A helper which use JEV to play ra2web(WannaFire Version)[王二火大]

利用 jev 刮削音乐元数据,风格,语言

Claude Opus 5.5 with the effort level re-decided every step by the TypeSafe Jev reflex — without breaking the prompt cache. CLI + Claude Code plugin.

Fast System-1 Decision, Arbitration & Safety Engine for Autonomous AI Agents (Powered by TypeSafe Jev)

Any chat model, any engine (SGLang, vLLM, transformers) as a Jev-compatible probability decision service: one prefill, one label token

ReflexBench — open benchmark and evaluation harness for System One models and typed decision engines

Signal-based LinkedIn outbound scoring for HeyReach, running on Jev (TypeSafe System One).

Semantic code and document search CLI — BM25 recall + TypeSafe Jev calibrated precision

Live sales-call copilot on TypeSafe Jev: per-utterance closing probability, signals and next-best-move; syncs to a real recorded call; tailored phrasing via Claude Haiku 4.5, verified by Jev

Clean, traceable state briefings for TypeSafe's Jev model

JevNext · More than Choice — A simple algorithm that equips any Jev-like model with numerical control.

Jev VAD: client-side turn detection for Agora Conversational AI — manual SoS/EoS and semantic barge-in judged by TypeSafe Jev

GPT teacher consults for DSH: stateless one-shot codex teachers behind a hard per-task consult budget, a read-only sandbox and a deterministic prefilter.

Selenium 2048 player powered by expectimax search and TypeSafe Jev, with portrait FFmpeg recording.

Rubric-Based Zero-Shot Classification Benchmark: Jev vs Claude Haiku 4.5 vs Claude Sonnet 5 vs OpenJev on rubric-conditioned classification, chained decision execution, and exam grading -- with full price tracking.

A governed execution layer between typed-decision models (TypeSafe Jev) and agent/AI tooling like Claude Code, MCP, and AI SDKs, turning probabilistic decisions into deterministic, policy-controlled, and auditable actions.

Chrome extension that puts unread Gmail first, critical on top, labeled by TypeSafe Jev.

Drive a phone with a model that never writes a word. TypeSafe's Jev picks each action, phone-use runs it on iOS and Android.

Lightning-fast AI supply chain security auditor using Jev (System 1 models). Catch typosquatting and malicious scripts in milliseconds.

Visually configure Jev questions, test them live, export ready-to-use code.

Evaluate TypeSafe AI's Jev (System One Model) on TMMLU+ v1.1 — four-way choice via the API's own response schema, 100% parse rate by construction

Benchmarks of MCP tool selection accuracy and latency. V2 evaluates 143 Woku tools with GPT-6 Luna API calls and Jev; includes Go/Rust cores, one TypeScript SDK and reproducible reports.

Small, fast, honest decision models. Open alternative to Jev: zero-shot, fit on your labels in seconds, calibrated with a coverage guarantee, Jev wire-compatible.

A bounded Jev System-1 browser tool for Pi, OpenCode V2, Amp, and MCP, powered by agent-browser

用 JEV（TypeSafe System One 判断模型）自动完成 B 站「硬核会员试炼」的单文件 Python 脚本：手机只扫码登录一次，100 题在电脑上自动答完，零第三方依赖。