Featured Jev Tools and Projects
Documentation
docs.typesafe.ai
Concepts, primitives, API, patterns, and SDK guides.
概念、问题原语、API、设计模式及 SDK 使用指南。
Introducing System One Models and Jev
typesafe.ai
Launch post covering the model interface, RLCD, published performance claims, demos, and caveats.
System One 与 Jev 的发布文章，介绍模型接口、RLCD、官方公布的性能主张、演示和注意事项。
typesafe-sdk-js
typesafe-ai/typesafe-sdk-js
Official TypeScript and JavaScript SDK with inferred answer types.
官方 TypeScript / JavaScript SDK，支持自动推导答案类型。
typesafe-sdk-python
typesafe-ai/typesafe-sdk-python
Official synchronous and asynchronous Python SDK.
官方 Python SDK，同时提供同步与异步调用。
system-one-adapter-python
typesafe-ai/system-one-adapter-python
Drop-in adapter for comparing the System One interface with LLM providers.
可直接接入的适配器，用于对比 System One 接口与其他大语言模型服务商。
skills
typesafe-ai/skills
Official agent skills for building and evaluating System One workflows.
用于构建和评估 System One 工作流的官方 Agent Skills。
Models and aliases
docs.typesafe.ai
Version IDs, current limits, pricing, and the distinction between stable and preview aliases.
模型版本标识、当前限制、价格，以及稳定版与预览版别名的区别。
Confidence
docs.typesafe.ai
Distinguishes Choice/Score confidence from answer probability; Noul has no separate confidence field.
区分 Choice / Score 的置信度与答案概率；Noul 不提供独立的置信度字段。
Jev 1.13 jaggedness
docs.typesafe.ai
First-party limitations covering arithmetic, dates, distractors, adversarial state, and structural inconsistencies.
官方能力局限说明：算术、日期、干扰信息、对抗性状态输入，以及输出结构上的不一致。
Cloudflare AI
developers.cloudflare.com
Provider-maintained `typesafe/jev` integration accepting state and typed questions.
由服务商维护的 `typesafe/jev` 接入，接受状态信息和带类型的问题。
OpenRouter
openrouter.ai
Provider listing for `typesafe/jev-1.13`, alongside the moving `typesafe/jev-latest` alias.
提供 `typesafe/jev-1.13`，同时提供指向版本会变化的 `typesafe/jev-latest` 别名。
Vercel AI Gateway
vercel.com
`typesafe-ai/jev` through AI SDK's experimental `evaluate` interface; its Boolean primitive corresponds to TypeSafe's Noul.
通过 AI SDK 实验性的 `evaluate` 接口调用 `typesafe-ai/jev`；其 Boolean 原语对应 TypeSafe 的 Noul。
advocaat
pithings/advocaat
Small type-safe client for asking Jev questions about datasets.
轻量、类型安全的客户端，用来向 Jev 提出有关数据集的问题。
jev
dannote/jev
Elixir/OTP client designed around GenServer replies and pattern matching.
围绕 GenServer 回复与模式匹配设计的 Elixir / OTP 客户端。
jev-axi
shiftynick/jev-axi
CLI for picking, rating, checking, ranking, triaging, and guarding from the shell.
在命令行中完成选择、评级、检查、排序、分流与防护的 CLI 工具。
jev-dsl
inanna-malick/jev-dsl
Early-alpha Haskell DSL that encodes typed question packets and decodes answers; HTTP transport is left to the caller.
处于早期 Alpha 阶段的 Haskell DSL，用于编码带类型的问题包和解码答案；HTTP 传输需由调用方实现。
jev-mcp
blakestone-x/jev-mcp
MCP server exposing classify, score, check, match, and screen tools.
MCP 服务器，提供分类、评分、检查、匹配与筛选工具。
jev-mcp
BYK/jev-mcp
An eval-first MCP server for Jev, that returns typed judgments (noul, choice, score) with probabilities instead of generated text.
以评估为先的 Jev MCP 服务器，返回带概率的类型化判断（noul、choice、score），而不是生成文本。
jev-shell-history
mrnugget/jev-shell-history
Ranks existing zsh history entries for inline completion; accepting a suggestion does not execute it.
对已有 zsh 历史命令排序，提供行内补全；接受建议不会直接执行命令。
Jevbridge
gamesonrblx/Jevbridge
ACP/MCP adapter for using Jev alongside coding and chat models.
ACP / MCP 适配器，让 Jev 与编程模型、对话模型协同使用。
jevclient
AboveColin/jevclient
Async Python client for typed Jev questions and probabilities.
异步 Python 客户端，用于提交带类型的 Jev 问题并获取概率。
jevr
simxnherrera/jevr
Native R client for typed questions and provider-independent answers through TypeSafe or OpenRouter.
原生 R 客户端，经 TypeSafe 或 OpenRouter 提交带类型的问题，并获得不依赖具体服务商的统一答案。
laravel-typesafe-jev
Butochnikov/laravel-typesafe-jev
Laravel integration with typed responses, async requests, and testing fakes.
Laravel 集成，支持类型化响应、异步请求和用于测试的模拟实现。
ruby_decision_model
obie/ruby_decision_model
Ruby client with standard-library transport for TypeSafe and OpenRouter decision endpoints.
Ruby 客户端，使用标准库传输层接入 TypeSafe 与 OpenRouter 的决策端点。
semdecide
sharziki/semdecide
Typed semantic decisions for Unix pipelines and CI.
为 Unix 管道与持续集成（CI）提供类型化语义决策。
typesafe-go
zhirschtritt/typesafe-go
Idiomatic Go SDK for the TypeSafe API.
符合 Go 语言惯用方式的 TypeSafe API SDK。
typesafe-mcp
itsmostafa/typesafe-mcp
MCP connector that gives agents access to Jev decisions.
MCP 连接器，让 Agent 可以调用 Jev 的决策能力。
typesafe-sdk-java
Premo-Cloud/typesafe-sdk-java
Community Java 17 client for Choice, Score, and Noul, with an optional Spring Boot starter.
社区维护的 Java 17 客户端，支持 Choice、Score、Noul，并提供可选的 Spring Boot starter。
zod-jev
jomatsu/zod-jev
Pairs local Zod shape validation with Jev semantic validation.
将本地 Zod 结构校验与 Jev 语义校验组合使用。
agent-router
nidhi-singh02/agent-router
Pre-release Herdr integration that filters eligible coding models by quota and policy before Jev ranks them.
尚未正式发布的 Herdr 集成：先按额度和策略筛选可用的编程模型，再由 Jev 排序。
blink
ellipsis-dev/blink
Navigates file and directory names with Jev-guided walkers to find codebase paths for a natural-language query.
借助 Jev 引导的遍历器浏览文件名与目录名，为自然语言查询定位代码库路径。
Canny
qkal/Canny
Evidence ledger that challenges unsupported "done" claims from coding agents.
为编程 Agent 建立证据台账，质疑缺少依据的“已完成”声明。
commit-miner
devanshbatham/commit-miner
Classifies Git diffs and commit messages into change types and candidate security-fix/CWE labels for inspection.
将 Git diff 与提交信息归类为变更类型及候选安全修复 / CWE 标签，供进一步检查。
foreman
thruwire/foreman
Software-factory supervisor that uses Jev to keep coding agents on task.
软件工厂的监督工具，利用 Jev 让编程 Agent 持续围绕任务执行。
is-malicious
luantak/is-malicious
Scans source, configuration, build, and CI files with Jev, then reports suspicious behavior and implicated lines before the code is run.
在代码运行前，用 Jev 扫描源码、配置、构建及 CI 文件，报告可疑行为与涉及的代码行。
jev-codex-router
0xNatoshi/jev-codex-router
Per-turn Codex model, reasoning, and speed-mode routing.
逐轮为 Codex 选择模型、推理配置与速度模式。
jev-guard
leepokai/jev-guard
Cross-agent tool-call risk scoring with allow, ask, and deny outcomes.
对不同 Agent 的工具调用进行风险评分，输出允许、询问或拒绝的结果。
jev-pref
doeixd/jev-pref
Linter that has Jev check code changes against project preferences from `jev-pref.json` and feeds findings back to coding agents.
让 Jev 按 `jev-pref.json` 中的项目偏好检查代码变更，并将发现的问题反馈给编程 Agent。
jev-review
devagrawal09/jev-review
Staged code-review workflow with a local dashboard.
分阶段的代码审查工作流，配有本地仪表盘。
jev-review MCP plugin
NiazMorshed2007/jev-review
Local-first continuous software-quality review for coding agents.
面向编程 Agent、优先在本地运行的持续软件质量审查 MCP 插件。
jev-router
gargpratyush/jev-router
Chooses a model for each fresh Claude Code or Codex turn while wrapping the existing CLI.
封装现有 CLI，在 Claude Code 或 Codex 每个新轮次开始时选择模型。
jevwire
Brainwires/jevwire
MCP tools, an embeddable decision library, and advisory or restrictive Claude Code hooks; judgments do not grant native permissions.
提供 MCP 工具、可嵌入的决策库，以及建议型或限制型 Claude Code hooks；模型判断不会授予宿主原生权限。
opencode-jev-orchestrator
aaronshaf/opencode-jev-orchestrator
Keeps an OpenCode parent model fixed and uses Jev difficulty judgments to delegate harder turns to temporary subagents.
保持 OpenCode 的父模型不变，由 Jev 判断任务难度，将较难的轮次委派给临时子 Agent。
perch
lakeday-org/perch
Semantic code linter that evaluates code units against configurable Jev questions.
语义代码检查器，使用可配置的 Jev 问题评估代码单元。
pi-jev
y0usaf/pi-jev
Measured tool-call gate and general typed decision layer for the Pi coding agent.
为 Pi 编程 Agent 提供经过测量的工具调用关卡和通用类型化决策层。
pi-warden
DevMortimer/pi-warden
Pi extension that judges rule compliance, risky actions, stuck loops, and completion claims; enforcement depends on the hook and policy.
Pi 扩展，用于判断规则遵守情况、风险操作、卡住的循环与完成声明；具体执行约束取决于 hook 和策略。
skillbox
kitze/skillbox
Self-hosted skill library with optional Jev relevance recommendations over an authorized catalog.
可自托管的 Skills 库；可选择使用 Jev，在已授权的目录中推荐相关技能。
skillranker
Dicklesworthstone/skillranker
Rust CLI that ranks agent skills against live session context and can abstain.
Rust CLI，根据当前会话上下文为 Agent Skills 排序，也可以放弃作出选择。
supercov
supercorp-ai/supercov
Scores source files so coding agents can prioritize code-quality work.
对源文件评分，帮助编程 Agent 确定代码质量改进的优先级。
wakegate
shitianfang/wakegate
Experimental gate where Jev decides whether a timer or incoming event is worth resuming a sleeping agent's LLM; code skips only when Jev is confident and always wakes on user messages, errors, and a skip limit.
实验性的唤醒关卡：由 Jev 判断定时器或传入事件是否值得恢复休眠 Agent 的大语言模型。代码仅在 Jev 有把握时跳过；用户消息、错误或达到跳过次数上限时始终唤醒。
fast-dev-compaction
leonaaardob/fast-dev-compaction
Codex port that restores Jev-selected verbatim history around native session compaction.
面向 Codex 的移植实现，在原生会话压缩前后恢复由 Jev 选中的逐字历史记录。
fast-jev-compaction
tamaratran/fast-jev-compaction
Claude Code plugin and library that score tool-call/result pairs for deletion or truncation while retaining selected text verbatim.
Claude Code 插件与库，对工具调用 / 结果对评分以决定删除或截断，同时原样保留选中的文本。
pi-fast-jev-compaction
joelhooks/pi-fast-jev-compaction
Pi extension that prunes stale tool history and leaves summary compaction to Pi when pruning is insufficient.
Pi 扩展，裁剪过时的工具历史；裁剪不足时，由 Pi 完成摘要式压缩。
pi-jev-compact
ilkerulusoy/pi-jev-compact
Selective, verbatim context compaction for Pi using Jev model.
使用 Jev 为 Pi 进行选择性的上下文压缩，原样保留选中的文本。
pi-jev-context
kevinpita/pi-jev-context
Opt-in Pi extension that filters older messages from model requests while preserving the original session history.
需主动启用的 Pi 扩展，从模型请求中筛除较旧的消息，同时保留原始会话历史。
winnow
GhalebDweikat/winnow
Judges tool results before admitting them into Claude Code context.
在工具结果进入 Claude Code 上下文前，先进行判断筛选。
yoshi
compozy/yoshi
Experimental Claude Code/Codex proxy that uses Jev to prune request context while preserving tool-call protocol structure.
实验性的 Claude Code / Codex 代理，用 Jev 裁剪请求上下文，同时保持工具调用协议的结构。
jev-browser-use
wy-coliney/jev-browser-use
Codex browser skill that uses Jev for navigation and target selection while Codex handles text entry and outcome verification.
Codex 浏览器 Skill：Jev 负责导航与目标选择，Codex 负责文字输入与结果验证。
jev-ultrafast
browser-use/jev-ultrafast
Browser agent where Jev selects an operation and compatible DOM target, and a separate LLM supplies typed text.
浏览器 Agent：由 Jev 选择操作及匹配的 DOM 目标，另一个大语言模型负责提供需要输入的文本。
jev-voice-browser
moritzkremb/jev-voice-browser
Maps partial speech transcripts to browser intents and observed targets, with code deciding whether to act, wait, or ask.
将部分语音转写映射为浏览器意图和已观察到的目标，再由代码决定执行、等待或询问。
mobile-jev
droidrun/mobile-jev
Android agent using Mobilerun observations and bounded Jev actions; execute mode controls a real device, while the published Uber demo stops before booking.
使用 Mobilerun 观察结果与受限 Jev 动作集的 Android Agent。执行模式会控制真实设备；公开的 Uber 演示在下单前停止。
typesafe-computer-use
awlevin/typesafe-computer-use
macOS computer-use experiment using OCR plus bounded Jev action selection.
macOS 电脑操作实验，结合 OCR 与受限的 Jev 动作选择。
duckdb-jev
colliber/duckdb-jev
DuckDB extension that exposes Jev judgments as SQL values with return types derived from the declared criteria.
DuckDB 扩展，将 Jev 判断作为 SQL 值返回；返回类型由声明的判断标准决定。
HA-Jev
AboveColin/HA-Jev
Home Assistant integration exposing typed answers as sensors, automation actions, and an Assist conversation agent.
Home Assistant 集成，将类型化答案提供为传感器、自动化动作和 Assist 对话 Agent。
jev-curate
AkashPriyadarshii/jev-curate
Streaming filter and scorer for Parquet and JSONL datasets.
为 Parquet 与 JSONL 数据集提供流式筛选与评分。
jev-reviewer
choxos/jev-reviewer
Research-document extraction aid where Jev selects and verifies source lines for verbatim quotes; findings require human review and are not clinical decisions.
研究文档提取辅助工具：Jev 选择并验证原始行，提取逐字引文。结果需要人工复核，不构成临床决策。
jev-search
superagents-lab/jev-search
Uses Jev to select search sources and rank Search1API results, returning source links and snippets.
使用 Jev 选择搜索来源并对 Search1API 结果排序，返回来源链接与摘要片段。
jev-trader
jarrodwatts/jev-trader
Kuru/Monad trading experiment with optional Jev buy/sell decisions; defaults to a mock model and dry-runs without a private key, but configured execution can place real orders.
Kuru / Monad 交易实验，可选用 Jev 作出买卖决策。默认使用模拟模型，无私钥时进行试运行；配置实际执行后可以下真实订单。
jev-tree
reachjalil/jev-tree
Recursive choice over taxonomies larger than Jev's direct option limit.
在超过 Jev 单次直接选项上限的分类体系中，递归完成选择。
jevlogs
reachjalil/jevlogs
OpenTelemetry log triage before more expensive analysis.
在调用更昂贵的分析之前，先对 OpenTelemetry 日志进行分流。
jevql
kylemclaren/jevql
psql-shaped client and Go/TypeScript/Python SDKs for vanilla Postgres where Jev makes Noul, Choice, and Score judgements about individual table rows after the plain SQL has run on the server, and the client applies the resulting filter, sort, or group.
面向原生 Postgres 的类 psql 客户端及 Go / TypeScript / Python SDK。普通 SQL 先在服务器执行，再由 Jev 对各行作 Noul、Choice、Score 判断，客户端据此筛选、排序或分组。
jevsql
EugeneBoondock/jevsql
SQL-like filtering, ranking, classification, and scoring with natural-language predicates.
使用自然语言判断条件，以类似 SQL 的方式进行筛选、排序、分类与评分。
llama-index-jev
WiktorB2004/llama-index-jev
LlamaIndex reranker and selector using Jev Score and Choice answers, with configurable confidence handling.
基于 Jev Score / Choice 的 LlamaIndex 重排序器与选择器，支持配置置信度处理方式。
n8n-nodes-typesafe-jev
n3ndor/n8n-nodes-typesafe-jev
Community n8n node for asking multiple typed questions over workflow state.
社区 n8n 节点，可针对工作流状态一次提出多个带类型的问题。
pg-jev
realZachi/pg-jev
PostgreSQL extension for semantic questions over table rows.
PostgreSQL 扩展，对数据表行提出语义问题。
pg_typesafe
giuliosmall/pg_typesafe
Pre-alpha PostgreSQL C extension exposing Choice, Noul, Score, and batched judgments from SQL.
Pre-alpha 阶段的 PostgreSQL C 扩展，通过 SQL 暴露 Choice、Noul、Score 与批量判断能力。
tax-doc-classifier
kyotofin/tax-doc-classifier
Classifies text-bearing PDF pages into IRS form and page-kind candidates with a confidence gate; document triage, not tax advice, and scanned pages need OCR.
带置信度关卡的文档分类器，将含文本的 PDF 页面归为候选 IRS 表单与页面类型。用于文档分流而非税务建议；扫描页需要 OCR。
tiershift
iamvatsalpatel/tiershift
Policy-bounded model routing for TypeScript and Python.
在明确的策略边界内，为 TypeScript 和 Python 应用进行模型路由。
typesafe-jev-workflow
GiesN/typesafe-jev-workflow
LangGraph email-intent workflow using a typed Jev choice.
基于 LangGraph 的邮件意图工作流，使用 Jev 的类型化选择。
heist-one
AbdelStark/heist-one
Browser stealth game where Jev judges guards while deterministic code owns the world.
浏览器潜行游戏：Jev 判断守卫行为，确定性代码负责管理游戏世界。
jev-drone
RomanSlack/jev-drone
Simulated MuJoCo quadrotor with Jev making slower tactical judgments from processed camera observations; deterministic code controls flight.
MuJoCo 四旋翼无人机模拟：Jev 根据处理后的相机观察作出较低频的战术判断，确定性代码负责飞行控制。
jev-experiments
dabit3/jev-experiments
Collection of inspectable Jev demos, including scripted support conversations with typed intent, escalation, and suggested-response decisions.
一组可检查实现的 Jev 演示，包括脚本化客服对话中的类型化意图识别、升级处理与建议回复决策。
JevPilot
standardagents/jevpilot
Three.js driving simulation where Jev chooses among candidate paths and speeds while local code handles vehicle dynamics and geometry.
Three.js 驾驶模拟：Jev 在候选路径与速度中选择，本地代码负责车辆动力学与几何计算。
JevScape
Skyvern-AI/jevscape
RuneBench-based RuneScape harness that maps Jev choices to a bounded game-action catalog and records tick-level results.
基于 RuneBench 的 RuneScape 测试框架，将 Jev 选择映射为受限游戏动作集，并记录逐 tick 结果。
killmyidea
monteduro/killmyidea
Startup-idea evaluator that chooses kill, fix, or ship.
创业点子评估器，在放弃（kill）、修改（fix）或推出（ship）之间作选择。
tsai-sc
phyous/tsai-sc
Original StarCraft shareware controlled with recorded Jev action probabilities.
通过 Jev 控制原版《星际争霸》共享软件版本，并记录动作概率。
typesafe-mario
fhshaik/typesafe-mario
Super Mario Bros. agent choosing actions from structured emulator state.
《超级马里奥兄弟》Agent，根据结构化模拟器状态选择动作。
typesafe-snake
sorrycc/typesafe-snake
Snake autoplayer with one typed decision per tick and code-generated legal moves.
贪吃蛇自动玩家，每个 tick 作一次类型化决策，合法动作由代码生成。
jevmeter
ChetasLua/jevmeter
Scores every sentence in a video and renders the result as an overlay.
为视频中的每句话评分，并将结果渲染为叠加图层。
Jevthoven
cocktailpeanut/jevthoven
Symbolic-music studio where Jev chooses plans, instruments, and bar patterns, and code renders editable music and MIDI.
符号音乐工作室：Jev 选择创作方案、乐器和小节模式，代码生成可编辑音乐与 MIDI。
SlidePilot
harshil1712/slidepilot
Experimental Slidev controller that judges speech transcripts for slide completion, with deterministic checks and manual navigation.
实验性的 Slidev 控制器，根据语音转写判断当前幻灯片是否讲完，并保留确定性检查与手动翻页。
Sponsor Skip
trungdq88/youtube-sponsor-detection
Finds sponsor reads in YouTube transcripts or transcribed audio while code owns timestamps and playback skipping.
从 YouTube 字幕或音频转写中识别赞助口播；时间戳和播放跳过由代码负责。
unclutter
kitze/unclutter
Browser extension that uses Jev to identify page clutter and saves reusable, reversible hiding rules.
浏览器扩展，使用 Jev 识别页面杂乱内容，保存可复用、可撤销的隐藏规则。
Vibe Check for X
RafalWilinski/vibecheck
Chrome extension that scores draft posts and reply context before posting; optional media descriptions come from a separate vision model.
Chrome 扩展，在发布前为推文草稿和回复上下文评分；可选的媒体描述由独立视觉模型提供。
Jev Visual
hr98w/jev-visual
Educational MLX/Qwen vision-language experiment sharing image context across candidate-scoring questions; its probabilities are not calibrated correctness estimates.
面向教学的 MLX / Qwen 视觉语言实验，在候选评分问题间共享图像上下文；输出概率不是经过校准的正确率估计。
Jevlike
vinnylarouge/jevlike
Trainable encoder and option-attention head for variable candidate sets, with separate visual game experiments.
为可变候选集合提供可训练的编码器与选项注意力头，另有独立的视觉游戏实验。
LitJev
zhengxuyu/litjev
Reproduction of Jev that turns any Qwen model into a fast decision model, serving the same `/v1/systemone` schema (Choice, Score, Noul) with no training and no generated answer text.
项目自述为 Jev 复现：无需训练、不生成答案文本，将 Qwen 模型转为快速决策模型，并提供相同的 `/v1/systemone` 接口结构（Choice、Score、Noul）。
jevmlx
bnsd55/jevmlx
Jev-style parallel constrained decisions for MLX models on Apple Silicon.
为 Apple Silicon 上的 MLX 模型提供 Jev 风格的并行受限决策。
kev
jaredpalmer/kev
Qwen2.5-0.5B adapter and decision head with training code, released weights, and parallel typed-question inference.
基于 Qwen2.5-0.5B 的适配器与决策头，提供训练代码、已发布权重以及并行的类型化问题推理。
NanoJev
TianyuCodings/NanoJev
Small parallel-decision model with dynamic candidates, a training pipeline, and recorded game comparisons that include shared code planning.
小型并行决策模型，支持动态候选项、训练流水线；已记录的游戏对比包含共享的代码规划逻辑。
openjev
zhihz/openjev
Local bilingual probability decisions from context, questions, and candidate answers.
根据上下文、问题与候选答案，在本地进行双语概率决策。
OpenJev (DiffusionGemma)
razorback16/openjev
Independent Jev-compatible server over DiffusionGemma/vLLM; the documented setup requires custom vLLM patches.
基于 DiffusionGemma / vLLM 的独立 Jev 兼容服务器；文档中的部署方案需要自定义 vLLM 补丁。
openjev-sglang
ekzhang/openjev-sglang
Jev-compatible API endpoint backed by open models and prefill-only inference.
由开放模型与仅预填充（prefill-only）推理支持的 Jev 兼容 API 端点。
openvons
genai-craft/openvons
Open decision layer for finite options across text, images, and Japanese voice commands.
面向有限选项的开放决策层，支持文本、图像与日语语音指令。
parallelConstraintDecoding
stephanj/parallelConstraintDecoding
Java and llama.cpp experiments in parallel constrained decoding.
使用 Java 与 llama.cpp 探索并行受限解码的实验。
reflex
kshetrajna12/reflex
Open-model decision engine with shared-state inference, isolated question branches, and a WebGPU demo; browser and Python configurations differ.
开放模型决策引擎，采用共享状态推理、隔离的问题分支，并提供 WebGPU 演示；浏览器与 Python 配置不同。
SemIf
TheoLeeCJ/SemIf
Formerly OpenJev: an independent study of typed option readout from frozen open models, with shared-prefix experiments and a WebGPU demo.
原名 OpenJev：研究如何从冻结的开放模型中读取类型化选项，包含共享前缀实验及 WebGPU 演示。
Simple Jev
featherless-ai/simple-jev
Transforms compatible open-model logits into typed decisions without a separately trained classifier head; model compatibility is constrained.
将兼容开放模型的 logits 转为类型化决策，无需单独训练分类头；模型兼容性有限制。
Verdict-open-jev
Heman10x-NGU/Verdict-open-jev
ModernBERT decision engine with calibrated uncertainty and a WebGPU playground.
基于 ModernBERT 的决策引擎，提供经过校准的不确定性与 WebGPU 体验环境。
Janus
FirasSX914/Janus
Measures when to use Jev versus other models and routes accordingly.
测量何时应使用 Jev、何时应使用其他模型，并据此进行路由。
jev-behavior-study
RINNECODER/jev-behavior-study
Independent synthetic-task study of Jev 1.13.0 framing sensitivity and failures, with raw responses and offline report checks.
对 Jev 1.13.0 问题表述敏感性与失败模式的独立合成任务研究，附原始响应和离线报告检查。
jev-benchmarks
AbdelStark/jev-benchmarks
Reproducible evaluation for calibration, selective risk, and latency.
用于校准、选择性风险与延迟的可复现评估。
jev-eval
4esv/jev-eval
Independent Jev versus GPT-5.6 Terra comparison on three labeled classification tasks, reporting accuracy, calibration, latency, and cost.
在三个带标签的分类任务上，独立对比 Jev 与 GPT-5.6 Terra，报告准确率、校准、延迟和成本。
jev-eval-agent
vinilana/jev-eval-agent
Compares LLM tool selection with Jev routing in a personal-assistant harness containing 100 mocked tools.
在包含 100 个模拟工具的个人助理测试框架中，对比大语言模型工具选择与 Jev 路由。
jev-korean-benchmark
mahlernim/jev-korean-benchmark
Small Korean/English sample study with recorded responses, including medical-text questions; not a clinical validation.
韩文 / 英文小样本研究，保留模型响应，包含医学文本问题；不构成临床验证。
jev-search-rerank-eval
zhuyansen/jev-search-rerank-eval
Chinese/English retrieval evaluation comparing Jev reranking with lexical, embedding, and fusion baselines, including judge-circularity analysis.
中英文检索评估，将 Jev 重排序与词法、向量嵌入及融合基线对比，并分析评判者循环论证问题。
jevcal
abhixhek/jevcal
Fits and drift-checks confidence thresholds against labeled data.
基于带标签的数据拟合置信度阈值，并检查其漂移。
typesafe-ai-benchmark
iammrduncan/typesafe-ai-benchmark
LLM gateway that mimics the System One output shape for comparison work.
模拟 System One 输出结构的大语言模型网关，用于对比研究。
Building with Jev
dbreunig/building-with-jev-skill
Community agent skill covering question design, state preparation, confidence thresholds, and debugging decisions.
社区 Agent Skill，介绍问题设计、状态准备、置信度阈值和决策调试。
Classifying RAG passages
docs.typesafe.ai
Official example of judging retrieved passages before passing them to an answering model.
官方示例：在将检索片段交给回答模型之前，先判断片段内容。
Date extraction
docs.typesafe.ai
Official pattern separating typed extraction from date validation and arithmetic in code.
官方模式：将类型化提取与代码中的日期校验、日期运算分开。
Double-checking citations
docs.typesafe.ai
Official example of checking whether source context supports a claim.
官方示例：检查来源上下文是否支持某项陈述。
Parallel questions
docs.typesafe.ai
Official worked example of evaluating many questions over shared state in one request.
官方完整示例：在一次请求中，针对共享状态并行评估多个问题。
Skill suggestion
docs.typesafe.ai
Official two-stage workflow that selects a skill and can reject the shortlist.
官方两阶段工作流：先选择 Skill，也可以拒绝整个候选列表。
awesome-jev-by-typesafe
Anil-matcha/awesome-jev-by-typesafe
Evidence-backed use cases, patterns, prompts, and starter code.
附有证据的用例、模式、提示词与起步代码。
awesome-jev
hellogumbo/awesome-jev
Large community directory with a searchable companion site.
大型社区导航目录，并提供可搜索的配套网站。
yibie/awesome-jev
yibie/awesome-jev
High-signal field guide organized by decision domain.
按决策领域组织的高信息密度实践指南。
awesome-typesafe
AbdelStark/awesome-typesafe
Broader TypeSafe and System One ecosystem list.
覆盖更广泛 TypeSafe 与 System One 生态的资源清单。
OmniJev/awesome-jev
OmniJev/awesome-jev
Papers, open reproductions, independent evaluations, and technical lineage.
收录论文、开放复现、独立评估与技术脉络。
awesome-jev-typesafe
valentynkit/awesome-jev-typesafe
CC0, awesome-lint clean, sorted by what you would install, with a short know-before-you-build section on the limits.
采用 CC0、通过 awesome-lint 检查，按准备安装的内容分类，并简要说明动手前需要了解的限制。
jev-go
Gaurav-Gosain/jev-go
Go client for TypeSafe's System One API and its model Jev: typed judgments and calibrated probabilities instead of generated text
面向 TypeSafe System One API 与 Jev 的 Go 客户端，返回类型化判断和概率，而不是生成文本；原清单对“校准概率”的描述不代表本站验证。
jev-go
Stumble/jev-go
An independent Go SDK for TypeSafe AI's Jev / System One API.
独立维护的 Go SDK，用于调用 TypeSafe AI 的 Jev / System One API。
swift-typesafe
ainame/swift-typesafe
Swift 6.4 SDK for TypeSafe AI, following the Python SDK's 0.6.0 API.
Swift 6.4 SDK，按原清单说明对齐 Python SDK 0.6.0 的 API；接入前请检查当前版本兼容性。
typesafe_sdk
nshkrdotcom/typesafe_sdk
typesafe_sdk is the Elixir SDK for TypeSafe AI and its first System One model, Jev.
用于 TypeSafe AI 及其首个 System One 模型 Jev 的 Elixir SDK。
typesafe_sdk_ex
vinnie357/typesafe_sdk_ex
Typesafe AI SDK in Elixir using Req
基于 Req 的 TypeSafe AI Elixir SDK。
typesafe-ai
Twister915/typesafe-ai
typesafe-ai brings TypeSafe's System One evaluation API into Rust as small, typed judgments that fit inside ordinary application code.
将 TypeSafe System One 评估 API 接入 Rust，把小型、类型化判断嵌入常规应用代码。
typesafe-ai-rs
gilljon/typesafe-ai-rs
An independent Rust client for the TypeSafe AI System One API, maintained at gilljon/typesafe-ai-rs.
独立维护的 TypeSafe AI System One API Rust 客户端，仓库为 gilljon/typesafe-ai-rs。
typesafe-rs
AbdelStark/typesafe-rs
Evaluate a state against named questions (noul, choice, score) and get one typed answer per question.
针对同一状态提交带名称的 Noul、Choice、Score 问题，并为每个问题返回类型化答案。
typesafe-sdk
joshmn/typesafe-sdk
A Ruby client for the TypeSafe System One API.
TypeSafe System One API 的 Ruby 客户端。
typesafe-sdk-go
Tangerg/typesafe-sdk-go
Go SDK for the TypeSafe AI API — typed questions in, probability distributions out.
TypeSafe AI API 的 Go SDK：输入类型化问题，返回概率分布。
typesafe-sdk-php
Butochnikov/typesafe-sdk-php
PHP client for the TypeSafe AI System One API.
TypeSafe AI System One API 的 PHP 客户端。
typesafeai-dotnet-sdk
saibimajdi/typesafeai-dotnet-sdk
Community .NET SDK for the TypeSafe AI System One API — typed noul, choice, and score questions with structured, confidence-scored answers. Not affiliated with TypeSafe AI.
社区维护的 .NET SDK，支持类型化 Noul、Choice、Score 问题与结构化答案；并非 TypeSafe 官方项目。原清单称其提供置信信息，但 Noul 本身没有独立 confidence 字段。
zio-typesafe-ai
jamesward/zio-typesafe-ai
A Scala 3 / ZIO library for TypeSafe AI's Jev / System One API — a "System One model" that answers typed, atomic questions about a piece of state instead of generating text.
Scala 3 / ZIO 库，用于调用 Jev / System One API，围绕给定状态回答类型化、原子问题，而不是生成文本。
ha-conversation-jev
luxus/ha-conversation-jev
Home Assistant custom conversation agent: Jev (TypeSafe System One, jev-latest) classifies an utterance, then either calls a light service (v0 fast path) or hands off to the SpaceXAI Grok conversation agent.
Home Assistant 自定义对话 Agent：用 Jev 对话语分类，再调用灯光服务（v0 快速路径），或转交 SpaceXAI Grok 对话 Agent。
hermes-jev
keeltrace/hermes-jev
It is built for narrow, typed judgments that should not require the main generative model to improvise an answer: routing, ranking, verification, multi-question assessment, tool gating, and context-value decisions.
为路由、排序、验证、多问题评估、工具调用把关及上下文价值判断提供狭窄、类型化决策，避免让主生成模型临场编写答案。
Jev MCP
jkudish/jev-mcp
MCP server that uses TypeSafe Jev for claim verification, content screening, and semantic candidate ranking.
使用 Jev 进行主张核验、内容筛查及语义候选排序的 MCP 服务。
jev-browser
Ying-Kai-Liao/jev-browser
Browser automation where an LLM plans and Jev (Typesafe System One) decides. Library, CLI and MCP server.
由大语言模型负责规划、Jev 负责判断的浏览器自动化项目，提供库、CLI 和 MCP 服务。
jev-ego
romaluev/jev-ego
TypeScript browser agent for ego lite — Jev Ultrafast's indexed action space, without Chrome, Playwright, or Browser Harness.
面向 ego lite 的 TypeScript 浏览器 Agent，采用 Jev Ultrafast 的索引动作空间，不依赖 Chrome、Playwright 或 Browser Harness。
jev-starter
hamakyo/jev-starter
Typed, policy-driven decision workflows on top of TypeSafe AI Jev: confidence routing, fallbacks, evaluation, and RAG patterns for TypeScript apps.
在 Jev 上构建类型化、策略驱动的决策流程，为 TypeScript 应用提供置信度路由、回退、评估和 RAG 模式。
ruby_llm-typesafe
kieranklaassen/ruby_llm-typesafe
TypeSafe structured-output provider for RubyLLM 2
为 RubyLLM 2 提供 TypeSafe 结构化输出的服务商适配。
typesafe-ai-rails
GenieRobot/typesafe-ai-rails
Community Rails integration for TypeSafe AI's System One API, built on the community typesafe-sdk Ruby gem.
基于社区 typesafe-sdk Ruby gem 构建的 Rails 集成，用于接入 TypeSafe AI System One API。
Jev Review Action
fatwang2/jev-review-action
Configurable GitHub Action using Jev to review directory submissions and classify pull requests, with evidence links and template-generated comments.
可配置的 GitHub Action：用 Jev 审查资源目录投稿并分类拉取请求，附带证据链接，由模板生成评论；最终合并由维护者决定。
jev-code
devagrawal09/jev-code
jev-code is a command-line toolkit that coding agents can delegate judgment-heavy work to.
供编程 Agent 委派大量判断类工作的命令行工具集。
jev-git
AkashPriyadarshii/jev-git
Sub-second Git pre-commit and pre-push reflex gate that screens staged diffs for secrets and destructive commands using TypeSafe Jev.
Git pre-commit / pre-push 检查工具，用 Jev 筛查暂存差异中的密钥与破坏性命令。原清单称其为“亚秒级”，本站未测量；不能替代其他安全检查。
jev-scout
AkashPriyadarshii/jev-scout
Zero-hallucination open-source repo and crate scout powered by TypeSafe AI Jev System One scoring.
用 Jev 评分发现开源仓库和 Rust crate。原清单使用“零幻觉”表述，本站保留原文供核对，不把这一宣传语作为可靠性保证。
jev-seo
AkashPriyadarshii/jev-seo
Zero-cost, agent-first SEO & Generative Engine Optimization (GEO) search radar CLI suite and MCP server powered by DuckDuckGo and TypeSafe Jev System One.
基于 DuckDuckGo 与 Jev 的 SEO / 生成式引擎优化（GEO）搜索雷达，提供面向 Agent 的 CLI 和 MCP 服务。原清单称“零成本”，实际 API 与服务费用需另行确认。
jev-superpowers
AkashPriyadarshii/jev-superpowers
Systematic software development framework for AI coding agents upgraded with TypeSafe Jev System One typed decisions, zero-hallucination package vetting, and completion gates.
面向 AI 编程 Agent 的系统化开发框架，加入 Jev 类型化决策、软件包核验和完成检查关卡。原清单对核验使用“零幻觉”描述，本站未验证该保证。
JevSeek
morcoan/JevSeek
Desktop and CLI coding agent that uses Jev to select the next tool from user intent and recorded execution results.
桌面与 CLI 编程 Agent，根据用户意图和已记录的执行结果，让 Jev 选择下一项工具。
Leanest
baronunread/leanest
Local-first test selector that uses TypeSafe Jev semantic judgments to decide which tests are safe to skip for a given code change, dropping straight into CI via its bundled GitHub Action.
本地优先的测试选择器，用 Jev 判断一次代码变更可以跳过哪些测试，附带 GitHub Action 以接入 CI；跳过测试的判断不构成正确性保证。
TypeSafe AI Playground
markjaquith/typesafe-ai-playground
Rust CLI with TypeSafe Jev experiments for PHI detection, code-comment review, tone analysis, and classification.
Rust 命令行实验集，展示受保护健康信息（PHI）检测、代码注释审查、语气分析及分类，不是临床验证工具。
hev reranker
hev/reranker
Python library using Jev Noul judgments to score candidate documents for query relevance, then sort or filter the results.
用 Jev Noul 判断给候选文档的查询相关性评分，再排序或过滤结果的 Python 库。
neo4jev
jexp/neo4jev
Neo4j graph navigation demo that uses TypeSafe Jev to select relationships and check goals during beam search.
Neo4j 图导航演示，在束搜索过程中让 Jev 选择关系并检查目标是否满足。
Jev-Trades
zadescoxp/Jev-Trades
A Next.js dashboard for live crypto market data and TypeSafe-powered paper trading.
Next.js 仪表盘，结合实时加密货币市场数据与 TypeSafe 驱动的模拟交易（paper trading）；不是投资建议或收益承诺。
typesafe-ai-playground
BunsDev/typesafe-ai-playground
A community playground for TypeSafe AI's Jev: edit classification experiments, compare A/B inputs, route conversations, extract document fields, inspect code-policy decisions, and explore games and simulations built around typed model outputs.
社区交互实验场：编辑分类实验、对比 A/B 输入、路由对话、提取文档字段、检查代码策略，并探索基于类型化模型输出的游戏与模拟。
UI Generator Instinct Jev
joevidev/ui-generator-instinct-jev
A demo app where the user describes a UI case in free text and Jev (TypeSafe's System One model) answers typed Choice/Noul/Score questions to pick and configure a real shadcn/ui component or page block, never generating code or copy.
根据用户描述的界面需求，由 Jev 回答 Choice / Noul / Score 问题，选择并配置既有 shadcn/ui 组件或页面区块；这是演示应用，不生成代码或文案。