8000 GitHub - rhc98/awesome-jev: Projects built on Jev (TypeSafe AI's System One model), curated by Jev itself. · GitHub
[go: up one dir, main page]

Skip to content

About

Projects built on Jev (TypeSafe AI's System One model), curated by Jev itself.

Topics

Resources

Contributing

Stars

1 star

Watchers

0 watching

Forks

Repository files navigation

Awesome Jev Awesome

Site Curated Judged Listed

Projects built on Jev, TypeSafe AI's System One model. Curated by Jev itself.

Every repository here was collected from GitHub, then judged by Jev in a single call: is it genuinely about Jev, which category, which decision pattern, and how substantial, documented, and novel it is. Code applies the policy; nothing below was hand-picked. The full index with probabilities lives on the site, along with the review queue of borderline cases.

This is an independent community project, not affiliated with TypeSafe AI. Jev and System One are TypeSafe AI product names.

Pipeline: GitHub search and community lists feed candidates into code, Jev judges each repository in one call, code applies the policy with human overrides, and the README and site are published daily.

Contents

Official

Maintained by TypeSafe AI.

SDKs and Clients

Language bindings, CLIs, and thin wrappers for the API.

  • jevi - Ask typed questions about a text and branch on the answer. A shell front end for TypeSafe's Jev.
  • jev-ultralightspeed - BRRRRRRRRRRRRRRRRRRRRRR.
  • solo_jev - Per-row semantic judgements over large relational tables on a decision model (Jev).
  • daf-jev - Composable Python toolkit for TypeSafe's Jev (System One) decision API — question builders, confidence gates, evaluator, calibration, CLI, MCP server, agent skill.
  • AnyDecisionModel - A Swift package for typed decisions from language models (probabilities, choices, and scores), with support for local MLX models and the TypeSafe Jev API.
  • ak-jev - Node.js bindings for TypeSafe's Jev — the System One model. Typed decisions, probabilities and confidence, with no text generation and no parsing.
  • ruby_decision_model - Ruby client for decision models such as Typesafe Jev.
  • laya - Run Laya, the open-source Jev-compatible System-1 decision model, from Node.js / TypeScript via ONNX Runtime.
  • jevkit - Fast Rust CLI for TypeSafe Jev: typed decisions, offline linting before you pay.
  • decision-gate - Rate and spend limits for code that calls Jev and other decision models.
  • jev-spec - ⚡ Catch spec drift on every commit: check your code against your Markdown specs with TypeSafe AI's Jev model.
  • jev-tree - Recursive Jev choice over a taxonomy. Select from more than 255 options without breaking TypeSafe Jev's choice cap.
  • ts-jev-cost-calculator - Unofficial CLI + Python estimator of tokens, cost and context limits for TypeSafe (System One / Jev) API requests. Not affiliated with TypeSafe.
  • typesafe-go - Zero-dependency Go SDK for TypeSafe's System One API (Jev). Typed questions in, calibrated probabilities out with static analyzers that catch bad question design at build time, a decision layer, batching and caching. Not affiliated with TypeSafe AI.
  • jev-cli - Small dependency-free CLI for TypeSafe Jev.

Integrations

Jev wired into frameworks, gateways, platforms, and databases.

  • pg-jev - Ask your PostgreSQL tables questions in plain language. A PostgreSQL extension powered by TypeSafe's Jev.
  • LLM2Jev - Turn local language models into Jev-style structured decision models. Get results from text and images with prefill alone—no token-by-token decoding required.
  • Jevstiller - Distill a repeated Jev classification task into a local model, on the fly — same answers, your hardware.
  • jevrag - A pluggable decision substrate for RAG pipelines; explicit, calibrated state → Decision → confidence → action gates, with Jev as the first swappable backend.
  • duckdb-jev - High-throughput, robust native DuckDB extension for batched and streaming TypeSafe/Jev classification, scoring, and semantic predicates from SQL.
  • jev4pg - Established strong JEV based natural language to SQL performance and well-covered semantic operators for PostgreSQL(PG) with optimized cost and latency. Reusable evidence, with native execution, parallel plans and explainable embeddings in development preview.
  • deepeval - The LLM Evaluation Framework.
  • Flide - 🔥🔥 The extremely fast inference engine for decision model.
  • dsh-jev-tools - Jev judgment, not generation: prune long tool output, screen fetched pages for injected instructions, and gate completion claims inside DeepSeek Harness.
  • rag-jev - Make room for useful evidence. Inspectable context selection for RAG, with Jev reranking and open benchmark studies.
  • ruby_llm-typesafe - TypeSafe structured-output provider for RubyLLM 2.
  • jev_nx - Open decision models as a Jev backend, running in-process on Nx.
  • jev-datafusion - DataFusion SQL functions for typed judgments. TypeSafe is one server.
  • jev - TypeSafe Jev for OTP: reply to Jev from a GenServer and pattern match on its answer.

Agent and Developer Tooling

Routers, guards, reviewers, skills, and MCP servers for coding agents.

  • hermes-jev-skills - Jev-powered model routing, memory, compaction, skill selection, computer and browser use for Hermes agents (also Claude Code and Codex).
  • jev-belay - Claude Code Stop hook that blocks an unverified done: reads the transcript for evidence, asks Jev once, fails open on everything else.
  • toolgate - A measured tool-call firewall for AI agents: allow, ask or deny before the agent acts. Claude Code mod and hook, MCP proxy. Built on TypeSafe Jev, measured on real sessions with a blind second labeler.
  • jev-browser - Browser automation where an LLM plans and Jev (Typesafe System One) decides. Library, CLI and MCP server.
  • Jevest - Automated PR review pipeline using Jev (TypeSafe AI) as a millisecond decision layer over an LLM reviewer.
  • Jevest - Automated PR review pipeline using Jev (TypeSafe AI) as a millisecond decision layer over an LLM reviewer.
  • pi-jev-auto-mode - Jev (TypeSafe System One) backed auto mode for the Pi coding agent: semantically auto-approves bash, write, and edit tool calls and fails closed when a decision cannot be made.
  • JevRouter - A lightweight Jev-powered router for models, tools, and subagents.
  • mu - This project (μ): most of what a coding agent decides isn't code. mu hands 40 kinds of yes/no calls to a small model: fewer wasted tokens, fewer approvals on autopilot, fewer runs that stop halfway. Built on pi.
  • jev-mcp - Fast, cheap, typed judgments from TypeSafe's Jev model, as MCP tools.
  • jev-judge-mcp - 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.
  • neurolink - The pipe layer of an AI nervous system — one interface connecting provider neurons to your application, across three inference types: generate, stream, and a calibrated decide (via TypeSafe Jev). MCP-native, voice (TTS/STT/realtime), RAG, memory, file processors. Powers Tara, Yama and Clairvoyance at Juspay.
  • jevgrep - Find code by asking what it does. A CLI for coding agents that uses Jev to discover relevant files and source context.

Applications

Products and features whose behavior depends on Jev decisions.

  • JevGuide - 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-Interaction - System-One-controlled full-duplex voice interaction: turn-taking as plain-language decisions judged by Jev.
  • TradingAgents - Multi-Agents LLM Financial Trading Framework.
  • classifier-dev - Zero-shot text classification over plain HTTP — no API key, no account. One Cloudflare Worker, a CLI, and an MCP server. https://classifier.dev.
  • fastbrowse - A fast browser agent: Jev picks each action from what is on the page, an LLM reads and plans, and every claim in an answer cites a quote from the page.
  • doc2mark - AI-powered Python library that converts any document (PDF, Word, Excel, PowerPoint, HTML) to clean Markdown while preserving complex tables and layouts using AI-Powered OCR technology.
  • openagents - Monorepo & docs.
  • jevgraph - Evidence-backed knowledge graph construction with typed Jev relation decisions.
  • jev-search - Search the web with TypeSafe's Jev: source selection, query understanding and relevance ranking. Built with Search1API.
  • jev-seo - Live SEO audit for any website from one homepage URL, judged by Jev. PDF, XLSX and Markdown reports.
  • beebots - Three AI trading bees on OKX, every decision by Jev. Paper trading by default. Not financial advice.
  • JEV-Paper-Radar - Turn research interests or paper links into a ranked reading list in Claude. Retrieve public abstracts, keep reasons and sources, and export HTML/CSV/JSON/RSS. Uploadable skill; no separate model API key. CLI and scheduled Jev/Claude workflows also included.
  • laya-opencv - OpenCV 与判定模型(Laya / Jev)结合的本地工作台:OpenCV 看图测量,硬规则管数字,模型下结论。用于水果采摘、杂草与害虫监测、钢管和纺织物质检。 | OpenCV + Laya/Jev decision models in a one-click local workbench: OpenCV measures, hard rules settle the numbers, the model gives the verdict. For agriculture and quality inspection.
  • jev-job-search - Find software internships and new-grad jobs, fill the application forms, check every answer on the page, and send the ones you approve. JEV decides, code acts, Claude writes. Runs on your machine with Claude Code.
  • yao - ✨ All your agents and workspaces in one place, on every device you own. Track tasks on a board, accessible from desktop, mobile, browser, or API. Self-hosted.

Games, Robotics, and Simulation

Jev in a control loop.

  • jev-card-agent - Auto paly with jev.
  • heist-one - Observable browser stealth game: Jev makes typed guard judgments while deterministic code owns the world.
  • jev-parallel-dispatch - Browser simulation for parallel Jev decisions with capacity-constrained assignment and retained evidence.
  • jev-dino - JEV action selection at 10 Hz meets original Chromium Dino physics. 13,724 points, 1080p60 recording, and auditable evidence.
  • jev-helper - A helper which use JEV to play ra2web(WannaFire Version)[王二火大].
  • embodied-jev - EmbodiedJev: MuJoCo robot decision workbench with MiniCPM5-2B, Jev and compatible model APIs.
  • board-games - Jev 棋馆:六种棋类对弈,TypeSafe 系统一模型 Jev 走子。零框架零构建纯静态,双击即玩。.
  • clash-jev - A Clash Royale bot with no trained policy: Jev (TypeSafe System One) makes every decision from the live game state.
  • jev-arena - Write a fighter in plain English. Jev pilots it in real time. PR-driven ladder, swappable brains.
  • jev-plays-pokemon-red - 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-jstris - Jev chooses Tetris placements; a local controller plays Jstris through keyboard input.
  • minesweeper-jev - Minesweeper where deterministic logic does the provable work and TypeSafe Jev is consulted only when the board forces a guess.
  • JEV-Arcade - Fifteen live AI games: TypeSafe Jev vs. Claude, GPT, OpenRouter and Ollama models.
  • jev-practice-speed - A WebGL demo where you play the card game Speed against a CPU whose brain is TypeSafe AI's Jev. The whole point of the app is to measure and show Jev's decision speed and decision accuracy in real time.

Research, Evals, and Reimplementations

Benchmarks, calibration studies, and open replicas.

  • decider - A family of System One-style models fine-tuned from Qwen3.5, designed for one-pass typed decisions with calibrated probabilities.
  • laya - Non-autoregressive System 1 decision engine. Typed choice, score and yes/no decisions over any text in a single forward pass, in 100+ languages, with a router that picks the right checkpoint per request.
  • kev - Jev-like family of decision models built on top of Qwen3.5/3.8 you can train and run on your own.
  • jevbench - This project by Benchmark Heaven: benchmark and leaderboard for AI decision models, measuring typed decision accuracy, calibration, latency and cost.
  • AnyJev - Turn any LLM into a Jev-style decision model: typed decisions, real probabilities, no training. (continue updating, welcome any issue and PR request).
  • RSI-Jev - Typed-decision models (noul / choice / score) trained by a self-improving loop of AI agents — checkpoints, the code that produced them, and every version that failed.
  • reflexbench - Open benchmark and evaluation harness for System One models and typed decision engines.
  • jev - Jevos: an open-source alternative to Jev and Microsoft-Decision-1. Yes/no decisions, choices and scores with probabilities, on your laptop's CPU.
  • open-spark-jev - Open-source, local decision models inspired by TypeSafe’s Jev and System One - built on Qwen3 for NVIDIA DGX Spark.
  • StartLux-Decision - Open decision models from 0.8B to 27B: a calibrated probability for every option in one pass, and a short thought only when unsure. Engine for vLLM, llama.cpp and MLX.
  • notjev - Super fast Jev like server, model agnostic, working with any OpenAI compatible endpoint.
  • vllm-jev - Native vLLM serving for Jev decision models.
  • sysone-bench - First independent head-to-head benchmark of System One decision models (Laya vs Jev) on byte-identical inputs.
  • privatemode-decisions-benchmark - Benchmarking Jev/Laya/GLM-5.3 Flash for System One-like tasks.

Learning

Tutorials, example galleries, and playgrounds.

  • Jev - Unofficial TypeSafe Jev showcase — System One decisions, not chat.
  • jev-lab - A hands-on lab for Jev, TypeSafe's System One model — AI decision trees, guardrails and routing with typed answers instead of text.
  • jev - Análise crítica e plano de aplicação do Jev em decisões estruturadas.
  • jev-builder - A browser form for building requests to TypeSafe's Jev: pick a template, fill in the blanks, copy the request. No JSON, no install, runs locally.
  • jev-starter - Typed, policy-driven decision workflows on top of TypeSafe AI Jev: confidence routing, fallbacks, evaluation, and RAG patterns for TypeScript apps.
  • jev-recipes - 248 small Jev decisions for JavaScript and TypeScript: route, rerank, gate, grade, compare, and label text for agents, RAG, support, code review, and music. Each returns a typed result with explicit uncertainty. Plug into Vercel AI SDK or LangChain agents, import one recipe, or run the CLI with JSON from any language.
  • jev-poc - A hands-on tour of Jev, TypeSafe's decisions model: twenty demos across seven shapes, with a cost-and-agreement assessment against a chat-model baseline.
  • jev-broadcast-lab - Testing Lab for Jev AI.
  • cookbook - Sixty notebook recipes for typed decisions with Jev: Choice, Noul and Score patterns from a first classification to closed-loop control.
  • jev-demo - Jev (TypeSafe System One) 学习与实测:概念文档 + 5 个可运行 demo + 可复现压测。实测 jev-1.13.0:扇出几乎免费,40 问与 1 问等延迟。.
  • decide-dont-generate - Code, labs and interactive tools for the book "Decide, Don't Generate": build the decision layer of AI agents with calibrated probabilities, cost-based thresholds and a free mock API. Runs in Python, no API key needed.
  • jev-decision-lab - A local lab for seeing what TypeSafe's Jev judgment model does on realistic business cases: typed answers, probabilities, policy in code, receipts.

Other Lists

Community-maintained lists and directories.

  • awesome-jev - Launch what you built on Jev, TypeSafe AI's System One decision model: a weekly board where makers launch and people vote, over an index of public Jev examples by decision pattern, with source citations, dated link checks and call-site text checks; runtime and performance are not independently tested here. EN/中文, JSON schema, platform matrix.
  • awesome-jev-survey - Awesome Jev: an evidence survey of Jev and Jev-like typed decision models — calibration, selective control and open implementations, with a searchable literature site.
  • what-is-jev - Independent, source-linked research on TypeSafe AI's Jev (System One), with 947 rubric-scored public repositories, recurring patterns, datasets, and bilingual documentation.
  • awesome-typesafe-jev - Awesome Jev: a source-backed field guide to TypeSafe's System One model, with SDKs, live demos, agent tools, and independent evaluations.
  • awesome-jev-projects - Awesome Jev: source-backed open-source ecosystem radar, plain-language project discovery, and automatic GitHub sync.
  • jev-in-the-wild - Real-world Jev use cases, open-source projects, benchmarks and criticism — what people actually build with TypeSafe AI's Jev, and where it fails. Machine-readable, updated daily.
  • awesome-jev-prompts - Question patterns, engine selection & ecosystem tracking for typed decision models — 63 entries / 10 categories, linter, engine cookbook, benchmark map, scenario cards, live leaderboard. EN/中文.
  • jev-lab - Hands-on research lab for TypeSafe's Jev (System One model): reproducible benchmarks of Noul/Choice/Score primitives, confidence gating, fan-out latency, agent control — plus a living audit of the Jev ecosystem.
  • jev-hub-catalog - Catalog and measurements of third-party specialists for the jev hub.
  • awesome-jev - A verified, community-maintained catalog of 997 open-source projects built with Jev.
  • jev-skills-market - Community Jev skill market and assistant for automation testing, finance research and voice workflows. Computer Use is an independent repository.
  • awesome-jev-verified - A curated list of open-source Jev projects where every entry links to the line of code that calls Jev, and every performance number comes from an independent 2,390-question benchmark.
  • awesome-jev-robustness - Tests, calibration audits and failure-mode studies of Jev (TypeSafe System One): jaggedness, consistency, injection, abstention.
  • < 7915 a href="https://github.com/wuyoscar/jev-skill">jev-skill - An awesome collection of Jev use cases, workflows, and agent skills.
  • building-with-typesafe-jev - Unofficial skill that teaches coding agents to build with TypeSafe AI's Jev: typed decisions, calibrated confidence, and prior art from 150+ community projects.

Other

Genuine Jev projects that fit no category above.

  • contextgem - Extract structured data and insights from documents using LLMs and decision models.
  • dgx-spark-qwen38 - This project is an open-source installer that turns an NVIDIA DGX Spark, or any GB10 machine (ASUS, Dell, HP, Lenovo, MSI, GIGABYTE, Acer), into a private AI server in one command: measured model lanes, OpenAI and Anthropic APIs, and a local cockpit.
  • alzeo - This project is an open-source installer that turns an NVIDIA DGX Spark, or any GB10 machine (ASUS, Dell, HP, Lenovo, MSI, GIGABYTE, Acer), into a private AI server in one command: measured model lanes, OpenAI and Anthropic APIs, and a local cockpit.
  • conjevture - Exact logic for uncertain answers. A small TypeScript library with explicit probability models and Jev converters.
  • ai-taskbar - Simples Project to show your use AI, with Sessions, etc.
  • jev-rankkit - Universal, type-safe reranking for Python objects, search, RAG, and agents.
  • JEVe - JEV-powered decision architecture for EVE Online.
  • dotfiles - Config files.
  • jev-quilt - JEV for quilt as understood output: cellular-first decision substrate — typed cells, hook-and-drop deltas, bookkeeper WAL, last-mile projection decoupled.
  • syllago-docs - Documentation site for Syllago—the package manager for AI coding tool content.
  • jevos-docker - Version-pinned Docker image for the jevos decision engine — OpenVINO INT8, CPU-only.
  • decision-datagen - Generates fine-tuning datasets for Jev-style decision models (choice/noul/score) from policy documents: shared states with typed questions, deterministic rule-derived gold answers, and calibrated teacher probability distributions — ready for System One / Laya-style training/fine-tuning pipelines.

How This List Is Made

Discover: GitHub repository search, code search for SDK usage, and links from other community lists. Enrich: README excerpt, file listing, manifest dependencies, and activity metadata. Judge: one Jev call per repository with 11 typed questions, gates as Noul, category and pattern as Choice, substance, docs, and novelty as Score. Curate: code applies thresholds and weights to the stored judgments, so changing policy never re-runs inference. Publish: this README and the site are generated from data/curated.json every day.

Question set v2, model jev-1.13.0. The calibration against human labels is on the how it works page.

Contributing

Missing project? Submit it with the repository URL. A bot checks the link, the next daily run judges it, and the verdict is posted back on the issue. This README carries the top 15 by composite in each category; everything else that passes the gate is on the site. Think Jev got one wrong? Say so and a bot drafts the data/overrides.yaml change for a maintainer to review. Do not edit this README directly; it is regenerated.

About

Projects built on Jev (TypeSafe AI's System One model), curated by Jev itself.

Topics

Resources

Contributing

Stars

1 star

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages

0