REPOSITORIES
Code.
All of it.
Every GitHub repository the index points at — official SDKs, integrations, replicas, benchmarks and demos — with stars, language and last push from GitHub.
404indexed addresses
1.19mstars · unique repos
21languages
402 unique repositories · 401 of 404 addresses with GitHub metadata · Activity & individual check dates → · repos.json
404 indexed addresses.
| Repository | Entry | Category | Lang | Stars | Pushed | License |
|---|---|---|---|---|---|---|
| Shubhamsaboo/awesome-llm-apps100+ AI Agents, Agent Skills and RAG Apps - Free and Open Source.Needle: semantic search inside webpages → | Needle: semantic search inside webpages | product | Python | 144k | Apache-2.0 | |
| TauricResearch/TradingAgentsTradingAgents: Multi-Agents LLM Financial Trading FrameworkTradingAgents: Jev social-post screening → | TradingAgents: Jev social-post screening | integration | Python | 122k | Apache-2.0 | |
| vllm-project/vllmA high-throughput and memory-efficient inference and serving engine for LLMsvLLM PR #57250: a Gemma-based Jev-like path → | vLLM PR #57250: a Gemma-based Jev-like path | repo | Python | 97k | Apache-2.0 | |
| virattt/ai-hedge-fundAn AI Hedge Fund TeamAI Hedge Fund: native Jev adapter → | AI Hedge Fund: native Jev adapter | integration | Python | 65k | MIT | |
| BerriAI/litellmThe fastest, litest AI Gateway. Rust core with Python SDK. Call 100+ LLM APIs in OpenAI (or native) format with cost tracking, guardrails, load balancing, and logging [Bedrock, Azure, OpenAI, Anthropic, OpenAI, VertexAI, vLLM, Nvidia NIM]LiteLLM: Jev-driven context compaction and pass-through → | LiteLLM: Jev-driven context compaction and pass-through | integration | Python | 63k | Other | |
| Wei-Shaw/sub2apiSub2API 一站式开源中转服务，让 Claude、Openai 、Gemini、Grok订阅统一接入，支持拼车共享，更高效分摊成本，原生工具无缝使用。Sub2API: native System One gateway → | Sub2API: native System One gateway | integration | Go | 46k | LGPL-3.0 | |
| tinyhumansai/openhumanThe fastest, cheapest, most efficient open-source agent harness. Run more than 500 agents on a $10 VPS.OpenHuman: Jev tool-search ranker → | OpenHuman: Jev tool-search ranker | integration | Rust | 43k | GPL-3.0 | |
| Yeachan-Heo/oh-my-claudecodeTeams-first Multi-agent orchestration for Claude CodeOh My ClaudeCode: opt-in Jev judgments → | Oh My ClaudeCode: opt-in Jev judgments | integration | TypeScript | 41k | MIT | |
| agentscope-ai/agentscopeBuild and run agents you can see, understand and trust.AgentScope: Jev classifier model → | AgentScope: Jev classifier model | integration | Python | 36k | Apache-2.0 | |
| NandhaKishorM/layaNon-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.Laya: the open decision model that predates Jev → | Laya: the open decision model that predates Jev | replica | Python | 35k | Apache-2.0 | |
| trycua/cuaScale computer-use 2.0 with open-source drivers, cross-OS sandboxes, and benchmarks for training, evaluation, and data generation.jev-use for computer use → | jev-use for computer use | integration | Rust | 32k | MIT | |
| ComposioHQ/composioComposio powers 1000+ toolkits, tool search, context management, authentication, and a sandboxed workbench to help you build AI agents that turn intent into action.Composio: TypeSafe tool decisions → | Composio: TypeSafe tool decisions | integration | TypeScript | 31k | MIT | |
| assafelovic/gpt-researcherAn autonomous agent that conducts deep research on any data using any LLM providersGPT Researcher: Jev context selection → | GPT Researcher: Jev context selection | integration | Python | 31k | Apache-2.0 | |
| browser-use/jev-ultrafastFastest and cheapest web agentJev Ultrafast (Browser Use) → | Jev Ultrafast (Browser Use) | repo | Python | 24k | MIT | |
| comet-ml/opikDebug, evaluate, and monitor your LLM applications, RAG systems, and agentic workflows with comprehensive tracing, automated evaluations, and production-ready dashboards.Opik: TypeSafe evaluation tracing → | Opik: TypeSafe evaluation tracing | integration | Python | 23k | Apache-2.0 | |
| vercel-labs/agent-desktopComputer use CLI for AI agentsAgent Desktop: reliable computer use through OS accessibility trees → | Agent Desktop: reliable computer use through OS accessibility trees | repo | Rust | 22k | Apache-2.0 | |
| milind-soni/OpenMausBotOpen-source Grok Bot alternative with a virtual machine that bots can useOpenMausBot: Jev agent selection → | OpenMausBot: Jev agent selection | integration | TypeScript | 21k | Apache-2.0 | |
| pydantic/pydantic-aiHow Python does AI. Agents, realtime voice, image generation, embeddings. Every model, every interface, typed end to end.Pydantic AI: first-class TypeSafeModel → | Pydantic AI: first-class TypeSafeModel | integration | Python | 21k | MIT | |
| vercel-labs/json-renderThe Generative UI frameworkjson-render: experimental Jev composition → | json-render: experimental Jev composition | integration | TypeScript | 20k | Apache-2.0 | |
| confident-ai/deepevalThe LLM Evaluation FrameworkDeepEval: structured Jev evaluation → | DeepEval: structured Jev evaluation | integration | Python | 19k | Apache-2.0 | |
| xerj-org/xerjAI search engine and MCP server for bots and agents. One Rust binary indexes code, docs, logs, JSONL/CSV and PDFs for BM25 full-text, vector and hybrid search, code search without grep, agent memory and RAG. Elasticsearch-compatible API; Corpus Hub serves hundreds of reference memory corpora.XERJ: Jev judgments over search results → | XERJ: Jev judgments over search results | integration | Rust | 19k | Apache-2.0 | |
| lidge-jun/opencodexUniversal provider proxy for OpenAI Codex & Claude Code — use any LLM (Claude, Gemini, Grok, DeepSeek, Ollama…) with Codex CLI, App, SDK, and Claude CodeOpenCodex: Jev model routing → | OpenCodex: Jev model routing | integration | TypeScript | 18k | MIT | |
| OpenByteInc/QuantDingerThe most complete open-source, self-hosted Agentic Trading OS—AI research, Python strategies, factor discovery, backtesting, optimization, paper/live multi-asset execution, risk controls, monitoring, MCP, and multi-tenant SaaS in one auditable stack.QuantDinger: Jev pre-trade decisions → | QuantDinger: Jev pre-trade decisions | integration | Python | 14k | Apache-2.0 | |
| run-llama/liteparseA fast, helpful, and open-source document parserDocJev: document classification and splitting with Jev → | DocJev: document classification and splitting with Jev | repo | Rust | 13k | Apache-2.0 | |
| Arize-ai/phoenixAI Observability & EvaluationOpenInference instrumentation for TypeSafe (Arize) → | OpenInference instrumentation for TypeSafe (Arize) | integration | Python | 12k | Other | |
| genspark-ai/genofficeFree, open-source AI Office suite: Docs, Sheets, Slides, PDF, Markdown and HTML editors with a built-in AI agent, plus a `genoffice` CLI and agent skill so Claude Code, Codex and Cursor can create and edit real .docx/.xlsx/.pptx files locally. Bring your own key. macOS, Windows & Linux.GenOffice: Jev file-search ranking → | GenOffice: Jev file-search ranking | integration | TypeScript | 12k | Apache-2.0 | |
| jaredpalmer/kevJev-like family of decision models built on top of Qwen3.5/3.8 you can train and run on your ownkev: a tiny Jev-like you can train on a MacBook → | kev: a tiny Jev-like you can train on a MacBook | replica | Python | 9.9k | Apache-2.0 | |
| jev-chat/jev-chat-jarvisThe chat decision assistant: before you reply, Jev reads the chat, judges intent and risk, and drafts replies you fill in with one tap. You press send. Android · Windows · macOS · iOS | 聊天辅助决策工具：先看懂对方再回复，发不发由你。jev-chat-jarvis: Android reply assistant → | jev-chat-jarvis: Android reply assistant | product | Kotlin | 8.3k | MIT | |
| tamaratran/fast-jev-compactionClaude Code plugin that replaces the compaction summary with Jev decisions: every tool call and result is scored in one fast request, stale ones are dropped or truncated, everything kept stays verbatim.Instant context compaction → | Instant context compaction | demo | TypeScript | 7.9k | MIT | |
| superdesigndev/tregOpenRouter for agent tools. Join community here: https://discord.gg/6mQYYfFMAnTreg + Jev: three workflows, $8K/month saved → | Treg + Jev: three workflows, $8K/month saved | integration | Python | 7.8k | Other | |
| 2FastLabs/agent-squadFlexible and powerful framework for managing multiple AI agents and handling complex conversationsAgent Squad: typed Jev routing → | Agent Squad: typed Jev routing | integration | Python | 7.8k | Apache-2.0 | |
| mizorewww/laya-mlxNative MLX runtime for Laya typed decision models — 7–14 ms short decisions on M3 Max. No text generation, PyTorch, or cloud API.laya-mlx / laya-coreml: Laya on Apple Silicon → | laya-mlx / laya-coreml: Laya on Apple Silicon | replica | Python | 7.1k | Apache-2.0 | |
| CopilotKit/openmuseA personal agent with a browser, terminal, files, and work that keeps going built with CopilotKit and AG-UI.OpenMuse: Jev decision controls → | OpenMuse: Jev decision controls | integration | TypeScript | 6.4k | MIT | |
| TheoLeeCJ/SemIf-OpenJevSemantic ifs from open models, on a 3090 at home. Independent; not affiliated with Jev or TypeSafe.Jev-like model on a home GPU → | Jev-like model on a home GPU | replica | Python | 4.9k | MIT | |
| TheoLeeCJ/SemIf-OpenJevSemantic ifs from open models, on a 3090 at home. Independent; not affiliated with Jev or TypeSafe.Jev-like model on a home GPU → | Jev-like model on a home GPU | replica | Python | 4.9k | MIT | |
| Asymptote-Labs/agent-beaconThe cross-harness, self-improving memory layer for AI agents. Join our community: https://discord.com/invite/zdNChS2fBuBeacon: cross-harness memory for coding agents → | Beacon: cross-harness memory for coding agents | repo | Go | 3.4k | MIT | |
| dzhng/jevgrepFind code by asking what it does. A CLI for coding agents that uses Jev to discover relevant files and source context.jevgrep: semantic code search for coding agents → | jevgrep: semantic code search for coding agents | integration | TypeScript | 3.2k | MIT | |
| Contrastive-LM/CLMContrastive Language Models (CLMs) are a new class of System One model trained with a contrastive learning objective that connects states and actions.CLM: an independent decision model → | CLM: an independent decision model | replica | Python | 3.2k | Apache-2.0 | |
| jarrodwatts/jev-traderOne AI trade decision every Monad block. Jev on Kuru MON-USDC.Jev Trader: one decision per Monad block → | Jev Trader: one decision per Monad block | demo | TypeScript | 3.1k | MIT | |
| rohitg00/pro-workflowClaude Code learns from your corrections: self-correcting memory that compounds over 50+ sessions. Context engineering, parallel worktrees, agent teams, and 17 battle-tested skills.Pro Workflow: advisory System One checks → | Pro Workflow: advisory System One checks | integration | JavaScript | 3.0k | — | |
| typesafe-ai/skillsAgent skills for building with TypeSafe's System One APIOfficial agent skills repository → | Official agent skills repository | repo | — | 2.9k | MIT | |
| zilliztech/memsearchA persistent, unified memory layer for all your AI agents (e.g. Claude Code, Codex, DSH), backed by Markdown and Milvus.MemSearch: Jev memory reranking → | MemSearch: Jev memory reranking | integration | Python | 2.9k | MIT | |
| HarleyCoops/Math-To-ManimCreate Epic Math and Physics Animations & Study Notes From Text and Images.Math-To-Manim: Jev checkpoint review → | Math-To-Manim: Jev checkpoint review | integration | Python | 2.9k | MIT | |
| TianyuCodings/NanoJevA nano replica of Jev: parallel decisions, dynamic candidates, and an end-to-end training pipeline.NanoJev → | NanoJev | replica | Python | 2.6k | MIT | |
| yibie/awesome-jevA curated list of public projects, integrations, and discussions built on Jev — TypeSafe AI's System One model for typed decisions.yibie/awesome-jev → | yibie/awesome-jev | directory | Python | 2.6k | — | |
| bespokelabsai/nimbleLocal typed decisions, contrastive data curation, and model evaluation.Bespoke Nimble: open data, open model, open recipe → | Bespoke Nimble: open data, open model, open recipe | replica | Python | 2.3k | — | |
| Paca-AI/pacaAI-native, free, open-source alternative to Jira, Trello, ClickUp & Monday. Built for Scrum teams where humans and AI agents collaborate as equals — on the same board, the same sprints, the same goals. Self-hosted. Fully customizable via config and plugins.Paca: Jev project and task routing → | Paca: Jev project and task routing | integration | Go | 2.1k | Apache-2.0 | |
| szczyglis-dev/py-gptDesktop AI Assistant powered by GPT, Gemini, Claude, Grok, Ollama, DeepSeek, Perplexity, and more - chat, agents, tools, MCP, plugins, RAG, vision, voice, image and video generation, speech synthesis and speech recognition, web search, memory, presets, canvas, assistants, themes, computer use. Windows, Linux, and Mac.PyGPT: inline System One decisions → | PyGPT: inline System One decisions | integration | Python | 2.0k | Other | |
| LLPhant/LLPhantLLPhant - A comprehensive PHP Generative AI Framework using OpenAI GPT 4. Inspired by LangchainLLPhant: PHP Jev classifier → | LLPhant: PHP Jev classifier | integration | PHP | 1.7k | MIT | |
| nicobailon/pi-mcp-adapterToken-efficient MCP adapter for Pi coding agentPi MCP Adapter: Jev semantic tool search → | Pi MCP Adapter: Jev semantic tool search | integration | TypeScript | 1.6k | MIT | |
| mizorewww/laya-coremlLocal Laya typed decisions on Apple Core ML and Neural Engine. Validated ports, ~5 ms short decisions on M3 Max, reproducible speed and energy benchmarks.laya-mlx / laya-coreml: Laya on Apple Silicon → | laya-mlx / laya-coreml: Laya on Apple Silicon | replica | Python | 1.6k | Apache-2.0 | |
| nokia-applied-research/AnyJevTurn any LLM into a Jev-style decision model: typed decisions, real probabilities, no training. (continue updating, welcome any issue and PR request)AnyJev: typed decisions from open models → | AnyJev: typed decisions from open models | replica | Python | 1.6k | Apache-2.0 | |
| agentconnect-md/agentconnectThe open-source, multi-agent alternative to Claude Tag. @ any agent, wherever work happens, they work alongside your team, learning as they go.AgentConnect: typed agent decisions → | AgentConnect: typed agent decisions | integration | TypeScript | 1.5k | Apache-2.0 | |
| autonomous-ai/openharnessThe ultimate harness for coding agents and beyond. All your agents. All your machines. One command center. Start with code, then follow your curiosity and build across disciplines: CAD, circuits, robots, games and music.OpenHarness: Jev Sheets question trials → | OpenHarness: Jev Sheets question trials | integration | TypeScript | 1.5k | MIT | |
| ollaya-dev/ollayaJev-compatible API server for open decision models. Ollama for decision models: pull and serve 29 model families (Laya, pplx-decider, Clef, Winnow, coba...) locally or self-hosted behind /v1/systemone. The TypeSafe SDK works unchanged: set its base URL and pick an open model. CLI, desktop app, Docker, MCP server, public bench.Ollaya: pull and serve open decision models locally → | Ollaya: pull and serve open decision models locally | repo | Rust | 1.5k | Apache-2.0 | |
| vinnylarouge/jevlikejevlike → | jevlike | replica | Python | 1.4k | MIT | |
| kerpopule/hermes-jev-skillsJev-powered model routing, memory, compaction, skill selection, computer and browser use for Hermes agents (also Claude Code and Codex)hermes-jev-skills: routing, memory, compaction, GUI clicks for Hermes → | hermes-jev-skills: routing, memory, compaction, GUI clicks for Hermes | integration | Python | 1.3k | MIT | |
| ENTERPILOT/GoModelAI gateway / AI control plane / AI proxy written in Go. Unified OpenAI-compatible and Anthropic-compatible API for OpenAI, Anthropic, Gemini, Groq, xAI, Ollama, vLLM and more. A LiteLLM alternative with observability, guardrails, streaming, cost tracking, intelligent routing, sticky sessions, failover, real-time logs and usage tracking. Prod ready.GoModel: native Jev gateway provider → | GoModel: native Jev gateway provider | integration | Go | 1.3k | MIT | |
| Arize-ai/openinferenceOpenTelemetry Instrumentation for AI ObservabilityOpenInference instrumentation for TypeSafe (Arize) → | OpenInference instrumentation for TypeSafe (Arize) | integration | Python | 1.3k | Apache-2.0 | |
| caliber-ai-org/ai-setupContinuously sync your AI setups with one command. Codebase tailor suited agent skills, MCPs and config files for Claude Code, Cursor, and Codex.Caliber ai-setup: optional Jev transcript compaction → | Caliber ai-setup: optional Jev transcript compaction | integration | TypeScript | 1.3k | MIT | |
| realZachi/pg-jevAsk your Postgres tables questions in plain language. A PostgreSQL extension powered by TypeSafe's Jev.pg-jev → | pg-jev | repo | Shell | 1.3k | Other | |
| awlevin/typesafe-computer-useComputer use for about $0.0002 a step: OCR the screen, classify the next action with TypeSafe, click. macOS.Computer use on Jev: OCR, classify, click → | Computer use on Jev: OCR, classify, click | demo | Python | 1.3k | MIT | |
| reticlehq/reticleAI agents can generate code, but still struggle to understand what they build. Reticle brings Jev-style machine-native runtime perception to web & desktop applications.Reticle: machine-native runtime perception for web and desktop apps → | Reticle: machine-native runtime perception for web and desktop apps | repo | TypeScript | 1.3k | Other | |
| anishfn/shapeshiftAn input that becomes what you mean: one text box that morphs into the right UI as you type. Powered by TypeSafe Jev, works offline.Shapeshift: text input that becomes a useful interface → | Shapeshift: text input that becomes a useful interface | demo | TypeScript | 1.1k | MIT | |
| TypeLLM/TypeLLMTypeLLM: LLMs with type-safe generationTypeAR: Jev-inspired type-safe decoding → | TypeAR: Jev-inspired type-safe decoding | replica | Python | 981 | Apache-2.0 | |
| TypeLLM/TypeLLMTypeLLM: LLMs with type-safe generationTypeAR: Jev-inspired type-safe decoding → | TypeAR: Jev-inspired type-safe decoding | replica | Python | 981 | Apache-2.0 | |
| feder-cr/jevjevos: an open-source alternative to Jev and Microsoft-Decision-1. Yes/no decisions, choices and scores with probabilities, on your laptop's CPU.jevos: TypeSafe-compatible decisions on a laptop CPU → | jevos: TypeSafe-compatible decisions on a laptop CPU | replica | C++ | 960 | MIT | |
| wfzyx/von archivedThe open-source System One decision model. Sub-15ms, non-autoregressive, local drop-in alternative to TypeSafe Jev.von: a CPU-only local decision model → | von: a CPU-only local decision model | replica | Python | 922 | Apache-2.0 | |
| Valen-Team/ValenOpen, lightweight multimodal decision models for text, images, and video. Train and run them yourself.Valen: independent multimodal decision model → | Valen: independent multimodal decision model | replica | Python | 909 | Apache-2.0 | |
| jev-chat/jev-chat-windowsJevChat-Windows：聊天窗口旁挂的回复辅助。窗口截图 + 本地离线 OCR 读对方消息 → Jev 判断意图 → 3 条候选一键填入，发送永远手动Jev Chat for Windows: ranked WeChat reply drafts → | Jev Chat for Windows: ranked WeChat reply drafts | product | Python | 884 | Other | |
| alisaitteke/photoshop-mcpMCP for Adobe Photoshop. 118 tools. Control from Cursor, Claude or custom LLMs.Photoshop MCP: Jev intent routing → | Photoshop MCP: Jev intent routing | integration | TypeScript | 849 | MIT | |
| githubnext/localjevLocalJev by GitHub Next: a Jev-compatible local API → | LocalJev by GitHub Next: a Jev-compatible local API | replica | TypeScript | 843 | MIT | |
| thruwire/foremanAn agent supervisor and software factory foreman powered by decision modelsForeman: a supervisor for coding agents → | Foreman: a supervisor for coding agents | post | Python | 823 | MIT | |
| kitfunso/hippo-memoryStop re-teaching your agent. Make your agent's memory work like a brain. Hippo is long-term memory for coding agents. It's a critical layer for your AI harness that connects across your different tools (Cursor, Claude Code, Codex). Local-first. Measured: about half the repeat mistakes of built-in memory alone, at the same cost.Hippo Memory: local memory with Jev ranking → | Hippo Memory: local memory with Jev ranking | integration | TypeScript | 792 | MIT | |
| razorback16/openjevOpen, Jev-compatible System One decision server on DiffusionGemmaopenjev: a Jev-compatible decision server on DiffusionGemma → | openjev: a Jev-compatible decision server on DiffusionGemma | replica | Python | 785 | Apache-2.0 | |
| chenxiachan/thoughtdagYour thinking deserves a map: an infinite canvas where LLM conversations grow into an editable thought graph. Wires are the context.ThoughtDAG: Jev conversation recall → | ThoughtDAG: Jev conversation recall | integration | TypeScript | 775 | MIT | |
| narumiruna/pi-extensionsA monorepo of Pi Coding Agent extensionsPi extensions: TypeSafe decisions and recall → | Pi extensions: TypeSafe decisions and recall | integration | TypeScript | 772 | MIT | |
| AnotiaWang/awesome-decision-modelsA curated list of decision models (System One / typed decision models): hosted APIs, open-weight models, runtimes, SDKs, applications, benchmarks, and papers.AnotiaWang/awesome-jev → | AnotiaWang/awesome-jev | directory | Python | 708 | CC0-1.0 | |
| devagrawal09/jev-reviewA staged code-review workflow and local dashboard built with TypeSafe Jev.jev-review: staged code review with a local panel → | jev-review: staged code review with a local panel | repo | TypeScript | 703 | MIT | |
| milind-soni/tiptour-macosOpen-Source fast local computer useTipTour: a local pointer companion for macOS → | TipTour: a local pointer companion for macOS | demo | Swift | 698 | Other | |
| nicobailon/surf-cliThe CLI for AI agents to control Chrome. Zero config, agent-agnostic, battle-tested.Surf CLI: bounded Jev browser commands → | Surf CLI: bounded Jev browser commands | integration | TypeScript | 687 | MIT | |
| gargpratyush/jev-routerRoute to the cheapest model in claude code for your task using jev-routerjev-router for Claude Code → | jev-router for Claude Code | repo | JavaScript | 662 | MIT | |
| AgentiLoop/AgentAgent! now supports macOS 14.6 or later, Apple Silicon and Intel, latest release. The Best AI agent for Mac: drives any app via Accessibility, codes/builds in Xcode, automates AppleScript, JXA, Swift and SMAppService shell (user/root). 23 LLM providers, local/cloud, plus on-device Apple Intelligence and Jev guard. Auto-Pilot, MCP, Avatar.Agent!: optional Jev action advisor → | Agent!: optional Jev action advisor | integration | Swift | 652 | Other | |
| featherless-ai/simple-jevTurn any open model into a classifier/jev endpointsimple-jev → | simple-jev | replica | Python | 652 | Apache-2.0 | |
| Sac-Y/Jev-cuJev-cu: the fastest computer use inside Codex → | Jev-cu: the fastest computer use inside Codex | integration | JavaScript | 631 | MIT | |
| coldteadotai/abideMake your coding agent abide by all your project rules abide: make your coding agent follow project rules → | abide: make your coding agent follow project rules | integration | TypeScript | 625 | MIT | |
| AbdelStark/awesome-typesafe-jevAwesome Jev: a source-backed field guide to TypeSafe's System One model, with SDKs, live demos, agent tools, and independent evaluations.awesome-typesafe → | awesome-typesafe | directory | HTML | 616 | MIT | |
| agentjido/req_llmComposable Elixir library for LLM interactions built on Req and FinchReqLLM: Elixir TypeSafe evaluation → | ReqLLM: Elixir TypeSafe evaluation | integration | Elixir | 607 | Apache-2.0 | |
| rmalde/minecraft-agentAstra planner and JEV controller for Minecraft, with native recording, tested routes, and run verification.minecraft-agent: Astra planner, Jev controller → | minecraft-agent: Astra planner, Jev controller | demo | JavaScript | 606 | — | |
| cobanov/awesome-jevA curated, source-backed list of projects built with Jev, TypeSafe AI's System One model for typed decisions.cobanov/awesome-jev → | cobanov/awesome-jev | directory | — | 592 | CC0-1.0 | |
| jkudish/jev-mcpFast, cheap, typed judgments from TypeSafe's Jev model, as MCP tools.jev-mcp → | jev-mcp | repo | JavaScript | 578 | MIT | |
| jerryjliu/docjevA very fast document classifier/splitter using Jev DocJev: document classification and splitting with Jev → | DocJev: document classification and splitting with Jev | repo | Python | 559 | Apache-2.0 | |
| superagents-lab/jev-searchSearch the web with TypeSafe's Jev: source selection, query understanding and relevance ranking. Built with Search1API.Jev Search → | Jev Search | repo | TypeScript | 559 | MIT | |
| TianyuCodings/JevHarnessLLM-authored task-specific Jev harnesses with optional full-trajectory reward reflection and GEPA evolution.JevHarness: authored decision pipelines → | JevHarness: authored decision pipelines | repo | Python | 559 | — | |
| jev-chat/jev-chat-jarvis-mac聊天悬浮窗助手（macOS）：屏幕感知 + 本地小模型判断意图与风险，按话术生成回复候选。纯只读。jev-chat-jarvis-mac: WeChat reply assistant → | jev-chat-jarvis-mac: WeChat reply assistant | product | Python | 529 | MIT | |
| kyotofin/tax-doc-classifierTax document page classifier built on Jev decisions. 100% strict accuracy across 261 IRS forms, ~$0.001 per page.Got Jev — now what? A real deployment list → | Got Jev — now what? A real deployment list | roundup | TypeScript | 519 | Apache-2.0 | |
| walidboulanouar/awesome-jev-use-casesAwesome list of TypeSafe AI Jev use cases: 74 demos ranked by likes, 150+ GitHub repos, limits, cost and API examples. CC0awesome-jev-use-cases: demos, patterns, limits, unbuilt ideas → | awesome-jev-use-cases: demos, patterns, limits, unbuilt ideas | directory | — | 491 | CC0-1.0 | |
| droidrun/mobile-jevmobile-jev: an Android agent on Jev → | mobile-jev: an Android agent on Jev | repo | JavaScript | 468 | MIT | |
| fhshaik/typesafe-marioA TypeSafe/Jev agent that plays Super Mario Bros. from structured emulator state.Jev plays Super Mario Bros from telemetry → | Jev plays Super Mario Bros from telemetry | demo | Python | 462 | — | |
| Zefan-Cai/Open-JevOpen-Jev: trained decision heads and an inspectable replay workbench → | Open-Jev: trained decision heads and an inspectable replay workbench | replica | Python | 450 | MIT | |
| moritzkremb/jev-voice-browserControl a real browser by voice. Jev (TypeSafe System One) decides intent + target in ~300 ms per spoken word; Playwright acts — often before you finish the sentence.Voice Browser: spoken commands to bounded browser actions → | Voice Browser: spoken commands to bounded browser actions | demo | JavaScript | 431 | MIT | |
| typesafe-ai/system-one-adapter-pythonDrop-in TypeSafeClient replacement backed by LLM APIsSystem One adapter for LLM APIs → | System One adapter for LLM APIs | repo | Python | 412 | MIT | |
| shhivv/third-handcomputer-use assistant w/ decision modelsthird-hand: computer use with no LLM at all → | third-hand: computer use with no LLM at all | demo | Swift | 401 | MIT | |
| malevrigns/agent-jevAgentJev-0.6B - a fast 'System One' decision model for AI Agents: feed it any unstructured state (diffs, traces, logs) and structured questions, get calibrated probability distributions back in one ~50ms forward pass. Zero output-token decoding.AgentJev: a small candidate-head model for agent decisions → | AgentJev: a small candidate-head model for agent decisions | replica | Python | 398 | Apache-2.0 | |
| kitze/unclutterWXT browser extension: Jev-powered page clutter removal with reusable template rules.Unclutter: Jev-powered clutter blocker → | Unclutter: Jev-powered clutter blocker | integration | TypeScript | 393 | MIT | |
| jkudish/jev-browserBrowser use using Typesafe's Jev modelJev Browser: Playwright agent through MCP, CLI or library → | Jev Browser: Playwright agent through MCP, CLI or library | repo | JavaScript | 368 | MIT | |
| itsmostafa/system-one-connectorSystem One MCP connector to evaluate anything fast and cheap. Give your AI agent direct access to models like: Jev, D1, CLM and Layatypesafe-mcp: Jev in Claude Code, Claude Desktop, Codex and Pi → | typesafe-mcp: Jev in Claude Code, Claude Desktop, Codex and Pi | integration | Go | 361 | MIT | |
| fstandhartinger/jevbenchJevBench by Benchmark Heaven: benchmark and leaderboard for AI decision models, measuring typed decision accuracy, calibration, latency and cost.JevBench: an independent benchmark for typed decision models → | JevBench: an independent benchmark for typed decision models | benchmark | Python | 339 | MIT | |
| FBddcz/embodied-jevEmbodiedJev: MuJoCo robot decision workbench with MiniCPM5-2B, Jev and compatible model APIsEmbodiedJev: an inspectable robot decision workbench → | EmbodiedJev: an inspectable robot decision workbench | demo | Python | 336 | MIT | |
| ekzhang/openjev-sglang archivedJev-compatible API endpoint based on open models (prefill-only)openjev-sglang → | openjev-sglang | replica | Python | 334 | — | |
| hr98w/jev-visualAn educational Jev-like visual inference experiment on Apple Silicon: shared context, direct candidate scoring, and local visual demos.jev-visual: Jev-like visual inference on Apple Silicon → | jev-visual: Jev-like visual inference on Apple Silicon | replica | Python | 330 | MIT | |
| logan-markewich/jeffA self-hosted drop-in replacement for TypeSafe's jev, powered by GliFormer.jeff: a self-hosted Jev API on GLiFormer → | jeff: a self-hosted Jev API on GLiFormer | replica | Python | 318 | MIT | |
| sutro-sh/jev-alignBuild calibrated AI Functions from human feedback using Jev and GEPA.jev-align: calibrate Jev to your decision criteria with GEPA → | jev-align: calibrate Jev to your decision criteria with GEPA | integration | Python | 317 | Apache-2.0 | |
| typesafe-ai/typesafe-sdk-pythonThe official Python library for the TypeSafe APItypesafe-sdk (Python) → | typesafe-sdk (Python) | repo | Python | 313 | MIT | |
| Heman10x-NGU/openJev-verdict-2.0Calibrated 151M Non-Autoregressive Decision Engine beating TypeSafe Jev & Laya on LocalLLaMA/typed-decisions (77.10% acc, 0.0636 Brier, 0.0144 ECE)openJev-verdict 2.0: a 151M calibrated decision engine → | openJev-verdict 2.0: a 151M calibrated decision engine | replica | Python | 306 | Other | |
| RomanSlack/jev-droneCamera-only autonomous drone in MuJoCo with a small judgment model (TypeSafe Jev) in the loop at 2.5Hzjev-drone: quadrotor obstacle course in MuJoCo → | jev-drone: quadrotor obstacle course in MuJoCo | demo | Python | 300 | MIT | |
| rokbenko/quackdOne CLI for all your robots. Connect them, command them, and let them work together, each with an LLM for a brain (Claude, OpenAI, Gemini, Grok, or local via Ollama or vLLM), VLAs for arms and decision LLMs (Jev, Laya, Kev). Drives Microduck, Open Duck Mini, LeRobot, XLeRobot, AlohaMini, ToddlerBot, or any ROS 2 base from a laptop, offboard.quackd: one CLI for all your robots, Jev for the cheap steps → | quackd: one CLI for all your robots, Jev for the cheap steps | integration | Python | 294 | Apache-2.0 | |
| typesafe-ai/typesafe-sdk-jsThe official TypeScript/JavaScript library for the TypeSafe API@typesafe-ai/sdk (TypeScript) → | @typesafe-ai/sdk (TypeScript) | repo | TypeScript | 285 | MIT | |
| monteduro/killmyideaDescribe your startup idea. Jev decides: kill it, fix it or ship it.Kill My Idea → | Kill My Idea | product | TypeScript | 279 | MIT | |
| 0xNatoshi/jev-codex-router archivedPer-turn model & reasoning routing for Codex, driven by Jev (TypeSafe System One): picks the model, thinking depth and speed mode for every turn.Jev Codex Router → | Jev Codex Router | repo | JavaScript | 274 | MIT | |
| langfuse/langfuse-docs🪢 Open source agent evals & observability: Trace, evaluate, and improve LLM applications with one open platform.Langfuse: using Jev for evals → | Langfuse: using Jev for evals | article | MDX | 253 | MIT | |
| standardagents/jevpilotA playable Three.js driving simulator with Jev-powered autopilotJevPilot: browser driving simulator → | JevPilot: browser driving simulator | demo | JavaScript | 238 | — | |
| aowang-ai/jev-tradeLive Jev trader on HyperliquidJev Trade: Hyperliquid trading bot → | Jev Trade: Hyperliquid trading bot | demo | TypeScript | 227 | Other | |
| hellogumbo/awesome-jevA community directory of projects built on Jev, TypeSafe AI's System One model.Awesome Jev → | Awesome Jev | directory | JavaScript | 226 | CC0-1.0 | |
| agent-labs-dev/fastbrowseA 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.Fastbrowse: browser tasks with typed decisions and citations → | Fastbrowse: browser tasks with typed decisions and citations | repo | Rust | 222 | MIT | |
| mmastrac/djev-sparkDiffusionGemma NVFP4 structured decisions on a DGX Spark: container recipedjev-spark: DiffusionGemma NVFP4 decisions on a DGX Spark → | djev-spark: DiffusionGemma NVFP4 decisions on a DGX Spark | replica | HTML | 220 | — | |
| ielab/llm-rankersDocument Ranking with Large Language Models.llm-rankers: checking Jev's calibration with TREC judgments → | llm-rankers: checking Jev's calibration with TREC judgments | benchmark | Python | 214 | Apache-2.0 | |
| jexp/neo4jevTypesafe.ai System One Model Jev navigating a Neo4j graph by using a classifier over neighbouring relationshipsneo4jev: graph navigation with Jev → | neo4jev: graph navigation with Jev | demo | Jupyter Notebook | 185 | MIT | |
| MaxBittker/runebenchBench mark long horizon problem solving in an MMOjevscape: RuneScape benchmark harness for Jev → | jevscape: RuneScape benchmark harness for Jev | benchmark | TypeScript | 175 | — | |
| kshetrajna12/reflexA small open decision model: state + typed questions -> calibrated probabilities. A Jev / System One re-creation on Qwen3.5.reflex: a small open decision model with a WebGPU build → | reflex: a small open decision model with a WebGPU build | replica | Python | 173 | MIT | |
| tamaratran/jev-prunerClaude Code plugin: trim long Bash output with TypeSafe Jev before the model sees itjev-pruner: trim tool output before the LLM sees it → | jev-pruner: trim tool output before the LLM sees it | repo | TypeScript | 168 | MIT | |
| allebee/jevk5JevK5: open-weight alternative to TypeSafe Jev. Typed decisions with probabilities in one forward pass; Apache-2.0 weights and code.JevK5: open decision weights with CUDA and GGUF runtimes → | JevK5: open decision weights with CUDA and GGUF runtimes | replica | Python | 167 | Apache-2.0 | |
| dbreunig/building-with-jev-skillA skill for writing and improving programs that call Jev, TypeSafe's System One modelBuilding with Jev skill → | Building with Jev skill | repo | — | 167 | — | |
| y0usaf/pi-jevTypeSafe Jev as a decision layer for the Pi coding agent: a measured tool-call gate plus jev_ask for typed, calibrated answerspi-jev (y0usaf): a measured tool-call gate for Pi → | pi-jev (y0usaf): a measured tool-call gate for Pi | integration | TypeScript | 165 | MIT | |
| zwliJay/jev-forgeAn open training and inference stack for Jev-style decision models. Train models to score dynamic candidate branches from a shared prefix, with support for high-cardinality choice, calibration, and fast batched inference.JevForge: decision-data synthesis, training and web-action scoring → | JevForge: decision-data synthesis, training and web-action scoring | replica | Python | 165 | Other | |
| leepokai/jev-guardAuto mode for every coding agent, built on Jev: risk-scores every tool call with session context (deny / ask / allow), flags prompt injection in results, checks skills and plugins. Claude Code, Codex, Copilot, Gemini, Cursor, pi, OpenCode, ACP.jev-guard: auto mode for every coding agent → | jev-guard: auto mode for every coding agent | integration | JavaScript | 153 | MIT | |
| kevinbadi/jev-voiceTalk to your Mac. Local whisper.cpp + one Jev (TypeSafe) call per command + macOS automation.Jev Voice: local speech and bounded Mac actions → | Jev Voice: local speech and bounded Mac actions | integration | Python | 137 | MIT | |
| daseinlabs/open-jevOpen Jev implementation with custom finetuningopen-jev (MLX, Apple silicon) → | open-jev (MLX, Apple silicon) | replica | Python | 136 | MIT | |
| qkal/CannyStops AI coding agents from claiming work is done without evidence. Deterministic hooks decide, TypeSafe's Jev advises. Append-only ledger, zero runtime dependencies.Canny: stop coding agents from claiming work is done without evidence → | Canny: stop coding agents from claiming work is done without evidence | integration | TypeScript | 135 | MIT | |
| SimpleJev/JevAnyOpen infrastructure for training, evaluating, and deploying System 1 decision models across language and multimodal backbones.JevAny: train and deploy decision models across backbones → | JevAny: train and deploy decision models across backbones | replica | Python | 135 | Apache-2.0 | |
| irfndi/prism-liquidity-agentAutonomous LP trading agent - auto-rebalancing with backtested strategies - Currently support Meteora DLMMprism-liquidity-agent: autonomous LP rebalancing on Meteora → | prism-liquidity-agent: autonomous LP rebalancing on Meteora | demo | TypeScript | 134 | MIT | |
| sdras/jev-webmcp-extensionA small extension that demos the combination of Jev x WebMCPJev × WebMCP: Chrome tool schema routing → | Jev × WebMCP: Chrome tool schema routing | integration | JavaScript | 134 | Apache-2.0 | |
| Ying-Kai-Liao/jev-browserBrowser automation where an LLM plans and Jev (Typesafe System One) decides. Library, CLI and MCP server.jev-browser: the LLM plans, Jev decides — library, CLI and MCP → | jev-browser: the LLM plans, Jev decides — library, CLI and MCP | integration | JavaScript | 134 | MIT | |
| Bodila51/grok-bot-jevConnect TypeSafe Jev to Grok Bot as a cheap decision layer - usage gates, skill template, examplesJev as a decision layer for Grok Bot → | Jev as a decision layer for Grok Bot | integration | Python | 133 | MIT | |
| savka777/jev-useSay it, and your Mac does it. A computer-use harness on Jev that reads the screen through Accessibility. Fast, no vision modeljev-use: a Mac computer-use harness without a vision model → | jev-use: a Mac computer-use harness without a vision model | demo | Swift | 131 | MIT | |
| devagrawal09/stanley-codeBounded TypeSafe Jev workflows for coding agents.Stanley: a Jev-first, self-evolving coding agent → | Stanley: a Jev-first, self-evolving coding agent | repo | TypeScript | 120 | MIT | |
| mrnugget/jev-shell-historyFish-style zsh history autosuggestions ranked by Jev (TypeSafe)jev-shell-history: fish-style zsh suggestions ranked by Jev → | jev-shell-history: fish-style zsh suggestions ranked by Jev | demo | TypeScript | 120 | — | |
| mustafakendiguzel/aso-workbenchApp Store Optimization on your Mac, with honest numbers. Rank tracking, Apple popularity, difficulty, AI listing writer and publishing to App Store Connect.ASO Workbench: Jev-assisted keyword analysis → | ASO Workbench: Jev-assisted keyword analysis | product | Swift | 120 | Other | |
| openlayer-ai/jevalsAgent evals and guardrails as Jev decisions: one request per trace, a fraction of a cent, fast enough for the agent loop. Runs locally with Kev or Laya.Openlayer jevals: batched decision-model evaluations and guardrails → | Openlayer jevals: batched decision-model evaluations and guardrails | repo | Python | 120 | MIT | |
| Heman10x-NGU/Verdict-open-jevNon-autoregressive decision engine on ModernBERT (151M) with calibrated uncertainty (RLCD), TypeSafe AI Jev benchmark audit, and in-browser WebGPU playgroundopenJev-verdict 2.0: a 151M calibrated decision engine → | openJev-verdict 2.0: a 151M calibrated decision engine | replica | Python | 118 | Other | |
| deepanwadhwa/OpenDecisionFree open source alternative for Jev that supports text, documents and images and runs at ~2GB of ram on a CPU.OpenDecision: local typed decisions with document evidence → | OpenDecision: local typed decisions with document evidence | replica | Python | 116 | Apache-2.0 | |
| trungdq88/youtube-sponsor-detectionDetect youtube sponsor segment with live audio and transcript powered by JevLive YouTube sponsor detection → | Live YouTube sponsor detection | repo | JavaScript | 116 | — | |
| GhalebDweikat/winnowA calibrated context sieve for Claude Code: every tool result is judged by a System One model before it enters context.winnow: a calibrated context sieve for Claude Code → | winnow: a calibrated context sieve for Claude Code | integration | Python | 115 | MIT | |
| AppitStudio/awesome-jevCurated Jev resources and runnable examples for typed AI decisions.AppitStudio awesome-jev: community project directory → | AppitStudio awesome-jev: community project directory | directory | Python | 114 | Other | |
| danielgshea/jev-as-a-judgeUsing Jev as an evaluator.Jev as a Judge: decision model vs LLM judges for evals → | Jev as a Judge: decision model vs LLM judges for evals | benchmark | Python | 114 | — | |
| ChetasLua/jevmeterPut a live Jev (TypeSafe) meter on any video: every sentence scored, rendered as a 16:9 editjevmeter: live fact-checking of spoken content → | jevmeter: live fact-checking of spoken content | demo | Python | 112 | MIT | |
| nassim-arifette/jevgrepJev-powered semantic code search for coding agents — find behavior across repositories via CLI or MCP, with exact source excerpts and line numbers.JevGrep: semantic source retrieval through CLI and MCP → | JevGrep: semantic source retrieval through CLI and MCP | repo | TypeScript | 112 | MIT | |
| AkashPriyadarshii/jev-curateHigh-throughput synthetic and pretraining dataset sifter for TypeSafe Jev. Rust streaming core, Parquet and JSONL I/O, typed Choice/Score/Noul judgments, speculative fan-out, 24.0 rows/sec measured. jev-curate: sift pretraining datasets at 1,500+ rows/sec → | jev-curate: sift pretraining datasets at 1,500+ rows/sec | integration | Rust | 111 | MIT | |
| ellipsis-dev/blinkCodebase search powered by Jev from @typesafe-aiblink: codebase search with Jev-scored paths → | blink: codebase search with Jev-scored paths | repo | TypeScript | 105 | — | |
| pithings/advocaatA small, type-safe client for asking AI questions about your data, powered by TypeSafe Jev.advocaat: a small type-safe client → | advocaat: a small type-safe client | integration | TypeScript | 101 | MIT | |
| sysevol-ai/CodeNibFind source context and trace calls across languages in Claude Code, Codex and other MCP agents.CodeNib: Jev reranking for code search → | CodeNib: Jev reranking for code search | integration | Python | 95 | Apache-2.0 | |
| wobsoriano/oxlint-plugin-jevoxlint-plugin-jev: plain-English rules over source code → | oxlint-plugin-jev: plain-English rules over source code | integration | TypeScript | 91 | MIT | |
| AboveColin/HA-JevAsk your house a question, get a number back. Home Assistant integration for TypeSafe Jev: typed answers as sensors, four actions for automations, and a conversation agent for Assist.HA-Jev: Home Assistant integration → | HA-Jev: Home Assistant integration | integration | Python | 89 | MIT | |
| Dimweaker/jev-liberoFine-grained robot control with Jev, physics previews, and configurable LIBERO tasks.Jev × LIBERO: robot decisions with local physics previews → | Jev × LIBERO: robot decisions with local physics previews | demo | Python | 87 | MIT | |
| sharziki/semdecideTyped semantic decisions for Unix pipelines and CI, powered by TypeSafe AI Jev.SemDecide: typed semantic decisions for Unix pipelines and CI → | SemDecide: typed semantic decisions for Unix pipelines and CI | integration | Python | 86 | MIT | |
| ikermoel/open-alternative-jevOpen-source alternative to TypeSafe's Jev: a System One style model layer that gives typed, calibrated decisions from any open-weights LLM in one forward pass (HF + vLLM), with honest benchmarksopen-alternative-jev: typed decisions from any open-weights LLM in one pass → | open-alternative-jev: typed decisions from any open-weights LLM in one pass | replica | Python | 80 | Apache-2.0 | |
| bnsd55/jevmlxJev-style parallel constrained decisions for any MLX model on Apple Silicon. Typed, schema-valid JSON in one forward pass.jevmlx: parallel constrained decisions for any MLX model → | jevmlx: parallel constrained decisions for any MLX model | replica | Python | 79 | MIT | |
| Reisenbug/TerraBlindA Terraria tModLoader mod where an AI fights the bosses: a model answers one question every 200ms, code turns it into keystrokes. Also a full code-only pipeline from fresh world to Wall of FleshTerraBlind: Jev boss-fight intent over Terraria controls → | TerraBlind: Jev boss-fight intent over Terraria controls | demo | C# | 79 | — | |
| kikoncuo/jevfireJEV-inspired parallel decisions for CUDA LLMs. One context, many decisions. vLLM API, game-agent examples, and reproducible benchmarks.JEVfire: parallel finite-choice decisions with vLLM → | JEVfire: parallel finite-choice decisions with vLLM | replica | JavaScript | 77 | MIT | |
| iamaamir/system-oneProvider-neutral System One runtime for TypeScript and PiSystem One for TypeScript: provider-neutral decisions → | System One for TypeScript: provider-neutral decisions | integration | TypeScript | 74 | — | |
| mizchi/jev-playgroundmizchi/jev-playground: MoonBit client and experiments → | mizchi/jev-playground: MoonBit client and experiments | repo | TypeScript | 73 | — | |
| iapp-technology/openthai-systemoneOpenThai-SystemOne: open Thai + English System One decision model (0.8B, 256-way slot head, Apache-2.0)OpenThai-SystemOne: an open Thai + English decision model → | OpenThai-SystemOne: an open Thai + English decision model | replica | Python | 69 | Apache-2.0 | |
| RenaGao/jev-dataopsAn open-source JEV-powered workbench for streaming data selection, quality evaluation, automatic LoRA training and held-out model evaluation.JEV DataOps: screen and evaluate training data → | JEV DataOps: screen and evaluate training data | integration | Python | 69 | MIT | |
| andududu/jeviewAn unofficial local visualizer for Jev (TypeSafe): a live view of every call your code makes. Not affiliated with TypeSafe AI.Jeview: a local proxy and inspector for Jev decisions → | Jeview: a local proxy and inspector for Jev decisions | repo | JavaScript | 68 | MIT | |
| burnigtm/jev-mcpMCP server that puts TypeSafe Jev on the coding loop in Cursor, Codex, and any MCP clientburnigtm Jev MCP: exact prepared call routing → | burnigtm Jev MCP: exact prepared call routing | integration | TypeScript | 68 | MIT | |
| achimala/jev-paintUse Jev to make art!jev-paint: use Jev to make art → | jev-paint: use Jev to make art | demo | JavaScript | 67 | MIT | |
| OmniJev/PlayJev🚀🚀 A 0.8B JEV-like multimodal model playing GUI games directly from raw pixels.PlayJev: a 0.8B visual decision model for ten games → | PlayJev: a 0.8B visual decision model for ten games | replica | JavaScript | 66 | Apache-2.0 | |
| EliaAlberti/jev-rulesJev picks which of your rules apply to each prompt, so Claude only sees the ones that matter.jev-rules for Claude Code → | jev-rules for Claude Code | repo | JavaScript | 65 | MIT | |
| lykycy123/RoboJEVTwo-stage JEV control of a Franka Panda in MuJoCoRoboJEV: two-stage control in MuJoCo → | RoboJEV: two-stage control in MuJoCo | demo | Python | 64 | Apache-2.0 | |
| r-ms/mini-jevmini-Jev: what a Jev-style typed-decision interface looks like on a frozen Qwen3-4B — read the option letter's logits instead of generating JSON. Preregistered experiment, results, teaching bench.WTF is Jev? Nine things people are already building → | WTF is Jev? Nine things people are already building | roundup | Python | 64 | MIT | |
| spring-ai-community/spring-ai-typesafeA Java SDK for the TypeSafe AI JEV API, & Spring AI TypeSafe integrations.Spring AI TypeSafe: Java SDK and Jev integrations → | Spring AI TypeSafe: Java SDK and Jev integrations | integration | Java | 64 | Apache-2.0 | |
| alvarobartt/sys1Blazing fast, self-hosted structured decisions for open models with a TypeSafe AI compatible API, written in Rust.sys1: a Rust server for open System One models → | sys1: a Rust server for open System One models | integration | Rust | 62 | Apache-2.0 | |
| Abhinavexists/levAn open System One decision modellev: a Qwen 4B decision model with dual readouts → | lev: a Qwen 4B decision model with dual readouts | replica | Python | 59 | Apache-2.0 | |
| obie/ruby_decision_modelRuby client for decision models such as Typesafe Jevruby_decision_model gem → | ruby_decision_model gem | integration | Ruby | 59 | MIT | |
| zostaff/agent-arenaagent-arena: a Jev bouncer for Degen Village → | agent-arena: a Jev bouncer for Degen Village | demo | TypeScript | 59 | MIT | |
| giuliosmall/pg_typesafePre-alpha PostgreSQL extension for TypeSafe AI (Jev) categorical classificationpg_typesafe: Jev from SQL (C extension) → | pg_typesafe: Jev from SQL (C extension) | integration | C | 57 | MIT | |
| mattt/AnyDecisionModelA Swift package for typed decisions from language models (probabilities, choices, and scores), with support for local MLX models and the TypeSafe Jev API.AnyDecisionModel: typed decisions from Swift sessions → | AnyDecisionModel: typed decisions from Swift sessions | integration | Swift | 57 | Apache-2.0 | |
| hotchpotch/jev-rerankerJev-powered relevance filtering and reranking for RAG in Python.jev-reranker: relevance filtering and reranking for RAG → | jev-reranker: relevance filtering and reranking for RAG | integration | Python | 56 | MIT | |
| skeptrunedev/jev-recruiterA Jev powered LinkedIn recruiting agent. Watch it browse relevant profiles, save links, and review evidence against your hiring brief.jev-recruiter: LinkedIn recruiting agent → | jev-recruiter: LinkedIn recruiting agent | demo | Python | 54 | MIT | |
| yusukebe/hono-jev-routerRoute HTTP requests by meaning. A semantic router for Hono powered by Jev.hono-jev-router → | hono-jev-router | integration | TypeScript | 54 | MIT | |
| h100envy/nervenerve: a Jev gate in front of an onchain trading agent → | nerve: a Jev gate in front of an onchain trading agent | demo | Python | 51 | MIT | |
| intikhab49/open-jev-typed-decision-engineOpen reproduction of TypeSafe Jev: a 150M typed decision engine (noul/choice/score in one non-autoregressive pass, calibrated confidence). 0.697 vs Jev's 0.727, 2.5x better calibrated, 4x faster, free. Trains on a Colab T4 in 30 min.open-jev: a 150M non-LLM typed decision engine → | open-jev: a 150M non-LLM typed decision engine | replica | Python | 50 | Apache-2.0 | |
| kbhuw/jev-siftClassify first. Read selectively. A portable agent plugin and MCP tool for batch text classification.jev-sift: classify first, read selectively → | jev-sift: classify first, read selectively | integration | JavaScript | 50 | — | |
| altryne/jevifyAn agent skill to discover TypeSafe Jev opportunities, design typed questions, and learn from recent community experiments.Jevify: an agent skill for typed-decision workflows → | Jevify: an agent skill for typed-decision workflows | integration | Python | 49 | MIT | |
| zhengxuyu/litjevTurn any off-the-shelf LLM into a Jev -like decision layerlitjev: /v1/systemone on top of a local Qwen via vLLM → | litjev: /v1/systemone on top of a local Qwen via vLLM | replica | Python | 48 | Apache-2.0 | |
| chy4pro/jev-for-chromeJev for Chrome: drives the tab you are looking at with TypeSafe Jev, a sub-second decision model. Community port of browser-use/jev-ultrafast, not affiliated with TypeSafe.Jev for Chrome: in-tab browser actions → | Jev for Chrome: in-tab browser actions | integration | TypeScript | 47 | MIT | |
| VGabriel45/polymarket-btc5m-jev-trading5m BTC Up/Down Polymarket trading agent using Typesafe Jev as the decision layer & TUIPolymarket 5-minute BTC Up/Down agent on Jev → | Polymarket 5-minute BTC Up/Down agent on Jev | demo | TypeScript | 47 | — | |
| mattn/go-jevGo SDK and CLI for TypeSafe Jev: typed decisions (yes/no, choice, score) from a modelgo-jev: Go client and UNIX decision pipelines → | go-jev: Go client and UNIX decision pipelines | integration | Go | 46 | MIT | |
| emrickgarrett/OneVOneJev1v1 Jev quickscope arena — Three.js + TypeSafe System OneOneVOneJev: a 1v1 quickscope arena on Three.js → | OneVOneJev: a 1v1 quickscope arena on Three.js | demo | TypeScript | 44 | — | |
| caiovicentino/eikosOpen, calibrated, single-pass typed-decision models (4B & 27B) for finance and tradingEikos: open typed decisions for financial rules → | Eikos: open typed decisions for financial rules | replica | Python | 43 | MIT | |
| JoshuaSP/open-jevTyped JSON inference with DiffusionGemma, with Every and Jev benchmark resultsopen-jev with DiffusionGemma → | open-jev with DiffusionGemma | replica | Python | 43 | MIT | |
| cclank/jevclipJev-powered video highlights and cited summaries from subtitles and scriptsjevclip: transcript-driven video highlights → | jevclip: transcript-driven video highlights | product | Python | 42 | MIT | |
| luantak/is-maliciousA codebase scanner that helps you not run malicous codeis-malicious: scan a repo before you run it → | is-malicious: scan a repo before you run it | integration | TypeScript | 41 | MIT | |
| mithalouni/system-one-openOpen replica of TypeSafe's Jev: typed calibrated decisions in one forward pass, on Gemma 4 E2B / Gemma 3 270M (Modal)system-one-open → | system-one-open | replica | Python | 41 | Other | |
| sgoedecke/system-oneBatched single-token choice inference for open language models, compatible with TypeSafesystem-one: batched single-token choice inference → | system-one: batched single-token choice inference | replica | Python | 40 | — | |
| yzfly/edgejev离线可用的本地类型化决策：4 核 CPU 单题 15.6ms。Local & offline Jev / System One inference on CPU — ONNX + INT8, no torch at runtime. 支持 laya / kev / PlayJevEdgeJev: export and quantize decision models for CPU → | EdgeJev: export and quantize decision models for CPU | integration | Python | 40 | Other | |
| Argos1111/jev_localReplicating Jev with a local LLMjev_local: TypeSafe-compatible adapter over LFM2.5 → | jev_local: TypeSafe-compatible adapter over LFM2.5 | replica | Python | 39 | — | |
| TypeSafeAI/jev-harnessA custom coding harness for TypeSafe AI's Jev: an LLM proposes, Jev answers narrow questions, code decides, every step leaves a receipt.Community jev-harness: proposal review with four questions → | Community jev-harness: proposal review with four questions | repo | TypeScript | 39 | MIT | |
| iammrduncan/inference-benchmarksThis is a LLM Gateway that mimics typesafe ai structured output. Like an imposter Jev.Needle 3 vs Jev: an 'imposter Jev' gateway benchmark → | Needle 3 vs Jev: an 'imposter Jev' gateway benchmark | benchmark | TypeScript | 38 | MIT | |
| AlbionaHoti/refgardenA spatial reference explorer for creators. Local Jev query choices, metadata highlights and source-linked collections.Refgarden: spatial reference search with bounded choices → | Refgarden: spatial reference search with bounded choices | demo | TypeScript | 37 | MIT | |
| bytelabs-oss/clash-jevA Clash Royale bot with no trained policy: Jev (TypeSafe System One) makes every decision from the live game stateclash-jev: live Clash Royale decisions and recorded replays → | clash-jev: live Clash Royale decisions and recorded replays | demo | Python | 37 | MIT | |
| DanRWilloughby/snifftestA prose linter that sniffs out AI writing tells. Zero dependencies, countable rules plus one judgment model.Sniff Test: a prose linter where Jev is the judge → | Sniff Test: a prose linter where Jev is the judge | repo | TypeScript | 37 | MIT | |
| lukaske/jev-doom-agentA browser-native Doom agent experiment with structured spatial state, composable AI controls, live decision telemetry, and a Chocolate Doom WebAssembly runtime.PROMPT FPS: browser-native Doom control → | PROMPT FPS: browser-native Doom control | demo | TypeScript | 36 | — | |
| colliber/duckdb-jevDuckDB extension: typed Jev answers as real SQL typesduckdb-jev → | duckdb-jev | integration | C++ | 33 | MIT | |
| jomatsu/pi-jev-auto-modeJev (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.pi-jev-auto-mode: typed tool permission gate → | pi-jev-auto-mode: typed tool permission gate | integration | TypeScript | 33 | MIT | |
| GodsBoy/jev-agent-skill-routerTyped, confidence-aware agent skill routing with TypeSafe Jev.Jev Agent Skill Router: bounded selection with abstention → | Jev Agent Skill Router: bounded selection with abstention | repo | Python | 32 | MIT | |
| mizzlelover/jev-hubJEV HUB · X 上关于 TypeSafe AI「系统一模型」Jev 的长文与演示视频聚合（保留原链与作者）｜ 谁是专家 出品JEV HUB: 216 demo videos and 214 long posts, categorised → | JEV HUB: 216 demo videos and 214 long posts, categorised | directory | CSS | 31 | Other | |
| Zaious/jev-capability-atlasIndependent, evidence-based map of when TypeSafe's Jev actually holds up vs. breaks down — real API-call receipts, not a leaderboard. 中文為主的雙語 repo。Jev capability atlas → | Jev capability atlas | benchmark | Python | 30 | MIT | |
| AbdelStark/jev-benchmarksProbability-aware evaluation for typed decision models: calibration, selective risk, latency, and reproducible benchmarks.Jev vs GLiNER2.5 benchmark → | Jev vs GLiNER2.5 benchmark | benchmark | Python | 28 | Apache-2.0 | |
| phyous/tsai-scTypeSafe Jev controls original StarCraft shareware through keyboard and mouse with recorded action probabilities.Jev plays StarCraft: a recorded Strongarm mission → | Jev plays StarCraft: a recorded Strongarm mission | demo | Python | 28 | MIT | |
| RiskAverseTech/toolgateA 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.toolgate: a tool-call firewall for agents → | toolgate: a tool-call firewall for agents | repo | TypeScript | 28 | MIT | |
| krzyzanowskim/TypeSafeTypeSafe SDK in SwiftTypeSafe SDK in Swift (krzyzanowskim) → | TypeSafe SDK in Swift (krzyzanowskim) | integration | Swift | 27 | MIT | |
| 0x7067/claude-jevClaude Code plugin: Jev for rule checks, verbatim compaction, and prompt routingclaude-jev: rule checks and verbatim compaction → | claude-jev: rule checks and verbatim compaction | integration | TypeScript | 26 | MIT | |
| evoke-build/evokeSoftware, by reflex. An open runtime that turns human intent into inspectable plans across small, composable programs, with explicit permissions and human approval before irreversible actions.Evoke: typed plans for installed reflex programs → | Evoke: typed plans for installed reflex programs | repo | Rust | 26 | Apache-2.0 | |
| marc2332/findmejev + fs = findmefindme: jev + fs → | findme: jev + fs | demo | Rust | 26 | — | |
| rorshopping/jev-on-a-laptopUnofficial study: Jev-style parallel typed decisions on stock 1.5B-8B models on an Apple Silicon laptop. Benchmarks, research notes, and a Hugging Face Space demo.jev-on-a-laptop: parallel constrained decoding on MLX → | jev-on-a-laptop: parallel constrained decoding on MLX | replica | Python | 26 | Other | |
| AgriciDaniel/gatekeeperRoutes 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.Gatekeeper: YAML rules for agent and skill selection → | Gatekeeper: YAML rules for agent and skill selection | integration | Python | 25 | MIT | |
| chopratejas/invalidateThe invalidation layer for AI memory. Every fact gets a lease; new evidence ends it. Built on TypeSafe Jev.invalidate: memory-lease invalidation with Jev → | invalidate: memory-lease invalidation with Jev | integration | Python | 25 | Apache-2.0 | |
| oso95/x-scannerChrome extension that labels every post you scroll past on X with typed Jev judgments and a live cost counterx-scanner: label X posts on five dimensions → | x-scanner: label X posts on five dimensions | integration | TypeScript | 25 | MIT | |
| yijunyu/jev-rsSystem One judgments (noul/choice/score) from any LLM in one prefill — a Rust, Jev-compatible /v1/systemone enginejev-rs: Rust System One scoring and evaluation engine → | jev-rs: Rust System One scoring and evaluation engine | replica | Rust | 25 | Apache-2.0 | |
| buberlo/dsh-jev archivedJev-powered decision layer for DeepSeek Harnessdsh-jev: Jev decision layer for DeepSeek Harness → | dsh-jev: Jev decision layer for DeepSeek Harness | integration | TypeScript | 24 | MIT | |
| khordoo/jev-reflex-autonomy-labReal-time drone swarm autonomy simulation using Jev for fast System 1 reflex decisions and collision avoidance, with optional System 2 reasoning for strategic guidanceReflex: a simulated drone autonomy lab → | Reflex: a simulated drone autonomy lab | demo | TypeScript | 24 | MIT | |
| kishida/llama.cppLLM inference in C/C++llama.cpp jev branch: local decisions on Qwen and Gemma → | llama.cpp jev branch: local decisions on Qwen and Gemma | repo | — | 24 | MIT | |
| braintrustdata/braintrust-sdk-pythonPython Tracing & Evals library for BraintrustBraintrust: Jev as a typed scorer for agent responses → | Braintrust: Jev as a typed scorer for agent responses | integration | Python | 22 | Apache-2.0 | |
| CommandCodeAI/cmd-mod-jev-nudgeCommand Code mod: nudges the agent to keep going when it stops with work left, judged by Jevcmd-mod-jev-nudge: decide when an agent should continue → | cmd-mod-jev-nudge: decide when an agent should continue | integration | TypeScript | 22 | MIT | |
| reachjalil/jevlogsOpen-source Jev log triage for OpenTelemetry. Score the signal before expensive LLM analysis.jevlogs: score log signal before the expensive LLM pass → | jevlogs: score log signal before the expensive LLM pass | integration | JavaScript | 21 | MIT | |
| fornhere/hafiza-osLinux, macOS ve Windows için kaynaklı ikinci beyin. Claude Code, Codex ve Antigravity adaptörleri; yerel Markdown kasa, ayrı hafıza incelemesi, isteğe bağlı Mem0/Jev.hafiza-os: a sourced second brain with optional Jev → | hafiza-os: a sourced second brain with optional Jev | integration | Python | 19 | MIT | |
| kieranklaassen/ruby_llm-typesafeTypeSafe structured-output provider for RubyLLM 2ruby_llm-typesafe → | ruby_llm-typesafe | integration | Ruby | 19 | MIT | |
| ruban-24/switchboard archivedOpen-source model and reasoning effort router for Claude Code and Codex, using Jev or experimental self-hosted Laya.switchboard: a cache-aware decision router for Claude Code and Codex → | switchboard: a cache-aware decision router for Claude Code and Codex | integration | TypeScript | 18 | Apache-2.0 | |
| ainame/swift-typesafeUnofficial Swift SDK for TypeSafeswift-typesafe: an unofficial Swift SDK → | swift-typesafe: an unofficial Swift SDK | integration | Swift | 17 | MIT | |
| sensahin/transcript-lensYouTube transkriptlerini anlamına göre keşfedin. Türkçe arayüz, Jev analizi, altyazı dışa aktarma ve Vercel kurulum rehberi.transcript-lens: explore YouTube transcripts by meaning → | transcript-lens: explore YouTube transcripts by meaning | demo | TypeScript | 17 | MIT | |
| Skyvern-AI/jevscapeRuneBench harness for TypeSafe's Jev: bounded action catalog, tick-mode controller and a live dashboardjevscape: RuneScape benchmark harness for Jev → | jevscape: RuneScape benchmark harness for Jev | benchmark | TypeScript | 17 | — | |
| mejiasd3v/pi-jev-routerAutomatic model routing for Pi using TypeSafe's Jev through Vercel AI Gatewaypi-jev-router: automatic model routing for Pi via Vercel AI Gateway → | pi-jev-router: automatic model routing for Pi via Vercel AI Gateway | integration | JavaScript | 16 | MIT | |
| syumai/jevyoumeanSemantic "Did you mean?" for any CLI — wraps commands and uses TypeSafe's Jev to match subcommand typos by intent, not edit distance.jevyoumean: semantic 'Did you mean?' for any CLI → | jevyoumean: semantic 'Did you mean?' for any CLI | demo | Go | 16 | MIT | |
| tyler-dot-earth/patdownBlock, steer, and "fuzzy lint" with Jev to make agents follow your rules and conventions. CLI, github action, pi package, claude extension, and more. Built with Effect + TypeScript.patdown: Markdown rule linting and agent write checks → | patdown: Markdown rule linting and agent write checks | repo | TypeScript | 16 | Other | |
| agent-chaperone/agent-chaperoneScreens an AI agent's tool calls before they run and tool results before the agent reads them. An MCP proxy plus a hooks adapter for a client's built-in tools.Agent Chaperone: MCP tool screening → | Agent Chaperone: MCP tool screening | integration | TypeScript | 15 | Apache-2.0 | |
| kavehmz/typesafe-playgroundInteractive experiments with TypeSafe Jev, from support routing to 3D driving simulations with real AI decisions and visible sensor inputs.Self-driving simulation with typed decisions → | Self-driving simulation with typed decisions | demo | JavaScript | 15 | — | |
| kylemclaren/jevqlSemantic SQL for Postgres, powered by Jevjevql: semantic predicates for vanilla Postgres → | jevql: semantic predicates for vanilla Postgres | integration | Go | 15 | MIT | |
| mohamadkhoshnava/ZeroNSFWBotZeroNSFWBot: optional Jev text moderation for Telegram → | ZeroNSFWBot: optional Jev text moderation for Telegram | integration | Rust | 15 | — | |
| prasanthj/duckdb-jevHigh-throughput, robust native DuckDB extension for batched and streaming TypeSafe/Jev classification, scoring, and semantic predicates from SQL.duckdb-jev (prasanthj): native, batched, cached → | duckdb-jev (prasanthj): native, batched, cached | integration | C++ | 15 | Apache-2.0 | |
| prismhq/jev-routerOpen-source LLM router that uses TypeSafe's Jev to pick a model, on top of LiteLLMjev-router on LiteLLM → | jev-router on LiteLLM | integration | Python | 15 | MIT | |
| Twister915/typesafe-aiTyped TypeSafe AI clients for Rust, with async and blocking backends and observable retries.typesafe-ai: Rust bindings for the System One API → | typesafe-ai: Rust bindings for the System One API | integration | Rust | 15 | Apache-2.0 | |
| arielweinberger/jev-autopilotThis demo uses Jev from TypeSafe AI to autonomously fly a drone in a random city from point A to point B, avoiding obstacles along the way. A trip costs $0.01.Jev Autopilot: typed drone stick controls in a generated city → | Jev Autopilot: typed drone stick controls in a generated city | demo | TypeScript | 14 | — | |
| bartlomein/okoBetter code search for coding agents: Claude Code, Codex, Cursor, OpenCode and Pi spend fewer tokens, tool calls and time.Oko: local code search and prompt-time retrieval → | Oko: local code search and prompt-time retrieval | repo | Rust | 14 | MIT | |
| forvela/jev-agent-browserFast, bounded browser agents powered by Jev and agent-browser — typed actions, research, classification, and safe orchestration.forvela Jev Agent Browser: bounded browser execution for parent agents → | forvela Jev Agent Browser: bounded browser execution for parent agents | integration | JavaScript | 14 | MIT | |
| TarunTomar122/jev-askable-armZero-shot English goals on a sim Franka. Jev chains hardcoded primitives.Askable Arm: English goals over a robot skill catalog → | Askable Arm: English goals over a robot skill catalog | demo | Python | 14 | MIT | |
| ajensenwaud/hermes-jev-pluginTypeSafe Jev (System One) decision tools for Hermes Agent: jev_check / jev_route / jev_score / jev_evaluatehermes-jev-plugin: decision tools for Hermes Agent → | hermes-jev-plugin: decision tools for Hermes Agent | integration | Python | 13 | MIT | |
| bohutang/siftChrome extension that labels every post on X (Substance · Humor · Chit-chat · Promo · Junk · AI-written) with TypeSafe Jev, and hides the ones you don't want.Sift: label every X post with Jev and hide the junk → | Sift: label every X post with Jev and hide the junk | product | JavaScript | 13 | MIT | |
| hyspacex/jev-routerRoute OpenAI-style chat requests to a model and reasoning effort, using TypeSafe's Jev decision model as the classifierhyspacex Jev Router: session model admission → | hyspacex Jev Router: session model admission | integration | Python | 13 | MIT | |
| jacks3tr/Jev-DesktopJev Desktop lets AI agents use Windows applications through MCP or the command line. Hand off a task and let Jev observe windows, choose controls, type supplied text, and navigate locally. Your agent gets the result or a request for help instead of handling every click.Jev Desktop: open-source computer use, ~10x faster than native → | Jev Desktop: open-source computer use, ~10x faster than native | demo | Python | 13 | MIT | |
| madeye/pi-jevJev-assisted file retrieval and request caching for faster Pi workflowspi-jev: Jev-assisted file retrieval and request caching for Pi → | pi-jev: Jev-assisted file retrieval and request caching for Pi | integration | TypeScript | 13 | MIT | |
| Nyarlathoteppppp/pi-heedRuntime constraints for the pi coding agent: checks every side-effecting tool call against what you said, before it runs. Powered by TypeSafe Jev.pi-heed: conversational constraints for Pi tool calls → | pi-heed: conversational constraints for Pi tool calls | integration | TypeScript | 13 | MIT | |
| tontoko/jev-browserOne grounded Jev/Playwright core: typed SDK, persistent CLI, and MCP server with native browser operations and deterministic assertions.tontoko Jev Browser: Playwright operations and semantic assertions → | tontoko Jev Browser: Playwright operations and semantic assertions | integration | JavaScript | 13 | Apache-2.0 | |
| abhixhek/jevcalStop guessing confidence thresholds: calibrate, threshold, and drift-check typed decision models (TypeSafe Jev) against an LLM teacher.jevcal: fit confidence thresholds and detect decision drift → | jevcal: fit confidence thresholds and detect decision drift | repo | Python | 12 | MIT | |
| DumoeDss/logjevLogJev: Jev-style choices from any model's first-token logprobs → | LogJev: Jev-style choices from any model's first-token logprobs | replica | TypeScript | 12 | — | |
| kushals256/jevcacheMorrowCache: open-source semantic cache for OpenAI-compatible LLM APIs. Local proxy that reuses answers for paraphrased prompts and AI agent retries. npm: @kushalicious/jevcachejevcache: skip the LLM call when Jev says same intent → | jevcache: skip the LLM call when Jev says same intent | integration | TypeScript | 12 | MIT | |
| anessbelbati/jev-rerank-benchCan a decision model beat dedicated rerankers? TypeSafe Jev vs Cohere Rerank 4 vs ZeroEntropy zerank-2 vs a chat-model baseline: 14 datasets, every raw API response, bootstrap ranges on every gap.Jev reranking: saved outputs across retrieval tasks and scoring strategies → | Jev reranking: saved outputs across retrieval tasks and scoring strategies | benchmark | Python | 11 | MIT | |
| GiesN/typesafe-jev-workflowLangGraph email workflow with Jev → | LangGraph email workflow with Jev | demo | Python | 11 | — | |
| limboinf/semantic-live-caption听写纸 · Real-time speech captions with live semantic annotation (key points / emotion / intent) — Confucius4-R2T2 + TypeSafe Jev + DeepSeekSemantic Live Caption: annotate a speech transcript → | Semantic Live Caption: annotate a speech transcript | integration | HTML | 11 | MIT | |
| matthewp/flue-jev-demoFlue agent routing with TypeSafe Jev through Cloudflare AI GatewayFlue agent + Jev via Cloudflare AI Gateway → | Flue agent + Jev via Cloudflare AI Gateway | demo | TypeScript | 11 | — | |
| rupeshpoojary9/poorjevOpen-source, local Jev alternative: a System One decision layer with provably calibrated confidence (ECE 0.170→0.071). Typed decisions, runs offline, no API key, no waitlist.poorjev: local typed decisions with calibration and abstention → | poorjev: local typed decisions with calibration and abstention | replica | Python | 11 | MIT | |
| zhuyansen/jev-search-rerank-evalDoes a TypeSafe Jev rerank beat embedding search? Graded relevance eval (9,831 pairs, 164 zh/en queries) over the Agent Skills Hub catalog, with the judge-circularity bias measured.Jev as a reranker: graded-relevance eval → | Jev as a reranker: graded-relevance eval | benchmark | Python | 11 | MIT | |
| larguesa/jev-searchExperimental semantic line search with TypeSafe Jev via OpenRouter. Python CLI with no runtime dependencies.larguesa jev-search: semantic line retrieval → | larguesa jev-search: semantic line retrieval | repo | Python | 10 | MIT | |
| mithril-lang/typed-decisionsJev-shaped typed-decision model (state + Choice/Score/Noul questions -> calibrated probabilities, one pass) on ModernBERT / DeBERTa / LLaDA-MoE, with measured latency, accuracy, calibration and training costopen-jev-deberta-v3-large (Kotoba Labs) → | open-jev-deberta-v3-large (Kotoba Labs) | replica | Python | 10 | Other | |
| sufianetaouil/everyAsk a yes/no question of every function in a codebase. Ranked answers in seconds, for cents. Grep whose pattern is a question, powered by TypeSafe Jev.every: ask a yes/no question of each source function → | every: ask a yes/no question of each source function | repo | Python | 10 | MIT | |
| AbdelStark/heist-oneObservable browser stealth game: Jev makes typed guard judgments while deterministic code owns the world.HEIST//ONE: typed guard judgments in a browser stealth game → | HEIST//ONE: typed guard judgments in a browser stealth game | demo | TypeScript | 9 | MIT | |
| jeffonelson/jev-bigquery-cloudrunClassify support tickets in BigQuery with Jev and Cloud RunJev decisions in BigQuery through Cloud Run → | Jev decisions in BigQuery through Cloud Run | integration | Python | 9 | — | |
| justhalfbit/dsh-plugin-jev-effort-selectorDeepSeek Harness (DSH) 推理等级自动选择插件：由 Jev System One 模型判断每条消息值多少思考量，按模型声明的等级自动推导档位，上下文信封让「继续」这类追问继承话题深度，低置信度向上取，任何失败都静默沿用原等级。 | Jev-driven reasoning effort per message: per-model ladders derived from what each model advertises, a fixed-size context envelope so follow-ups inherit topic depth, ties break upward, silent fallback on every failure path.DSH Jev Effort Selector → | DSH Jev Effort Selector | integration | JavaScript | 9 | MIT | |
| scienthoon/luceLuce: a recipe for calibrated decision models — a sentence about your task in, a small model that answers typed questions with honest probabilities out (init → synth → train → eval → serve)Luce: train and calibrate a decision model for your task → | Luce: train and calibrate a decision model for your task | replica | Python | 9 | Apache-2.0 | |
| spoonnotfound/soupbaseJev x 海龟汤Soupbase: bilingual lateral-thinking puzzles judged by Jev → | Soupbase: bilingual lateral-thinking puzzles judged by Jev | repo | TypeScript | 9 | MIT | |
| aaronshaf/opencode-jev-orchestratorKeeps OpenCode on a cheap sticky model for warm cache; Jev escalates hard turns to stronger subagents.opencode-jev-orchestrator: cheap sticky model, Jev escalates hard turns → | opencode-jev-orchestrator: cheap sticky model, Jev escalates hard turns | integration | TypeScript | 8 | MIT | |
| jammaru/jev-lab100 AI NPCs live in a tiny town. Jev chooses the next action; the world writes the story.Jev Lab: a 100-NPC town and a shogi opponent → | Jev Lab: a 100-NPC town and a shogi opponent | demo | TypeScript | 8 | MIT | |
| ranjan2829/AskJevAskJev — Jev autopilot for any website + guard on irreversible clicks (TypeSafe System One, not Claude)Got Jev — now what? A real deployment list → | Got Jev — now what? A real deployment list | roundup | TypeScript | 8 | MIT | |
| romeromarcelo/jev-retrievalSemantic code and document search CLI — BM25 recall + TypeSafe Jev calibrated precisionjevr: BM25 recall and Jev reranking for code and documents → | jevr: BM25 recall and Jev reranking for code and documents | repo | Rust | 8 | Apache-2.0 | |
| WiktorB2004/llama-index-jevLlamaIndex reranker + router powered by TypeSafe Jev - typed scores/choices, cheaper than LLM-as-judge.LlamaIndex reranker and selectors for Jev → | LlamaIndex reranker and selectors for Jev | integration | Python | 8 | MIT | |
| 0xtrou/rubikjevChallenge the Jev's intelligence in Rubik Cube puzzlesJev solves the cube → | Jev solves the cube | demo | TypeScript | 7 | — | |
| akash-kamat/system-one-gemmaOpen-source Jev-style System One decision model. Gemma 3 270M with a scoring head — fast, calibrated decisions in a single forward pass. No text generation. Inspired by TypeSafe.ai's Jev.system-one-gemma → | system-one-gemma | replica | Python | 7 | — | |
| arthurfiorette/jev-playwrightJev-powered Playwright test selectionJev Playwright: change-aware browser test selection → | Jev Playwright: change-aware browser test selection | integration | TypeScript | 7 | MIT | |
| chris-wozniczek/jev-voice-controlControl your Mac by voice. Speech → Jev (TypeSafe AI System One model) typed decisions → macOS actions. Menu-bar Swift app.Jev Voice: typed decisions for Mac voice control → | Jev Voice: typed decisions for Mac voice control | product | Swift | 7 | MIT | |
| enoyola/jev-grand-prixAn F1 racing game where TypeSafe's Jev picks the racing line and the pedals, and learns each corner's limit between lapsJev Grand Prix: racing lines and pedals → | Jev Grand Prix: racing lines and pedals | demo | JavaScript | 7 | MIT | |
| finetuningsingh/intelliprompterA teleprompter of talking points that checks each one off as you cover it, using TypeSafe's Jev Score questionsIntelliPrompter: check off talking points as you speak → | IntelliPrompter: check off talking points as you speak | demo | HTML | 7 | MIT | |
| i-priyanshuverma/laravel-jevlaravel-jev: Jev::is / Jev::choose inside Laravel → | laravel-jev: Jev::is / Jev::choose inside Laravel | integration | PHP | 7 | MIT | |
| legostin/jev-mcpMCP server that lets Claude Code, Codex & AI agents drive a real Chrome browser with JEV (TypeSafe System One): ~100 ms typed decisions, calibrated confidence, human-in-the-loop questions, safety gates and a debug trace.legostin Jev MCP: Chrome control and decision traces → | legostin Jev MCP: Chrome control and decision traces | integration | TypeScript | 7 | MIT | |
| satviksinha/jev-model-routerModel router for Claude Code using JevJev Model Router: turn-level model and effort hooks → | Jev Model Router: turn-level model and effort hooks | integration | TypeScript | 7 | MIT | |
| 0x7067/jev-browseBrowser automation with Jev (TypeSafe) as decision modeljev-browse: Jev-driven Chrome automation → | jev-browse: Jev-driven Chrome automation | repo | JavaScript | 6 | MIT | |
| deyna256/langchain-skill-routerPer-turn skill selection for LangChain and deepagents agents: a fast judge picks the few skills a turn needs, so a catalog of hundreds stays out of the prompt.LangChain Skill Router: per-turn skill selection with Jev → | LangChain Skill Router: per-turn skill selection with Jev | integration | Python | 6 | MIT | |
| kataras/jevA Go client for the TypeSafe AI's System One API and its model, Jev.kataras/jev: a Go client for System One → | kataras/jev: a Go client for System One | integration | Go | 6 | MIT | |
| lbotinelly/jev-little-airwaysA show-and-tell capability study for Jev, TypeSafe's System One decision model.Little Airways: aircraft and tower judgments in a toy airspace → | Little Airways: aircraft and tower judgments in a toy airspace | demo | HTML | 6 | MIT | |
| paulsmith/computer-use-jevmacOS computer use driven by Jev (TypeSafe System One) as the decision makerComputer Use Jev: macOS Accessibility automation → | Computer Use Jev: macOS Accessibility automation | integration | Go | 6 | MIT | |
| Stumble/jev-goCommunity Go SDK for TypeSafe AI Jev / System OneStumble jev-go: Go SDK and interactive CLI → | Stumble jev-go: Go SDK and interactive CLI | integration | Go | 6 | MIT | |
| xafold/jev-routerjev-router: automatic model and effort switching for Claude Codexafold Jev Router: cache-aware model routing for Claude requests → | xafold Jev Router: cache-aware model routing for Claude requests | integration | Rust | 6 | — | |
| Arpit-Khandelwal/jev-linkedin-slop-filterSlams a BAIT, CORP or BRAG stamp onto LinkedIn engagement-bait, judged live by Jev (TypeSafe System One).jev-linkedin-slop-filter: label LinkedIn posts → | jev-linkedin-slop-filter: label LinkedIn posts | integration | JavaScript | 5 | MIT | |
| atarikcaliskan/jevball22 Jev models, one ball: a 3D football match where every player is its own Jev (TypeSafe AI System One) decision. Watch, or take over the number 9.JevBall: 22 football players with inspectable decisions → | JevBall: 22 football players with inspectable decisions | demo | JavaScript | 5 | Other | |
| jamesward/zio-typesafe-aizio-typesafe-ai: a Scala ZIO client → | zio-typesafe-ai: a Scala ZIO client | integration | Scala | 5 | Apache-2.0 | |
| mandu5/jevcompatA testable spec and conformance suite for Jev-compatible (TypeSafe System One) API servers, with a normalising proxy, a reference mock and a GitHub Action.jevcompat: System One API conformance suite → | jevcompat: System One API conformance suite | repo | Python | 5 | MIT | |
| MANISH007700/tab-bouncerChrome extension that closes the tabs you don't need, judged by TypeSafe's Jev in one call. Tell it what you're doing; it shows the rest the door.Tab Bouncer: close the tabs you don't need, judged in one call → | Tab Bouncer: close the tabs you don't need, judged in one call | product | JavaScript | 5 | MIT | |
| Peu77/JevFindFast semantic code search powered by Jev. Find the relevant files, line ranges, and snippetsJevFind: two-stage semantic code finder → | JevFind: two-stage semantic code finder | repo | Rust | 5 | MIT | |
| Reindeer-AI/pi-jev-guardCheck Pi code edits against repository Markdown rules with TypeSafe Jevpi-jev-guard: check Pi code edits against repo rules → | pi-jev-guard: check Pi code edits against repo rules | integration | TypeScript | 5 | — | |
| zeredy879/minojevDecisions, not tokens: minojev reads calibrated, typed probability distributions straight from hidden states in one forward pass — zero output tokens, fully reproducible on a laptop CPU.minojev: a trained decision head on a frozen language model → | minojev: a trained decision head on a frozen language model | replica | Python | 5 | MIT | |
| adtyavrdhn/pydantic-jev-examplesPydantic AI capabilities made stronger with Jev: small runnable demos, one file eachPydantic AI: first-class TypeSafeModel → | Pydantic AI: first-class TypeSafeModel | integration | Python | 4 | — | |
| caiovicentino/jev-shieldSemantic MCP firewall powered by Jev — screens every tool call, tool result, and tool description with calibrated System One verification. 94% block recall, 0 false positives, ~$0.00002/check.jev-shield: semantic MCP firewall → | jev-shield: semantic MCP firewall | repo | JavaScript | 4 | MIT | |
| cappuch/openjev.cppllama.cpp fork for jev-like modelsopenjev.cpp: a llama.cpp fork for Jev-like inference → | openjev.cpp: a llama.cpp fork for Jev-like inference | replica | C++ | 4 | MIT | |
| FogMoe/necroAbandoned Qwen3.5-0.8B LoRA fine-tuning experiments for Jev-like typed judgments, with datasets, adapters, evaluations, and a full retrospective.necro: an abandoned Qwen3.5-0.8B LoRA replica, with the retrospective → | necro: an abandoned Qwen3.5-0.8B LoRA replica, with the retrospective | replica | Python | 4 | Apache-2.0 | |
| KantaHayashiAI/jev-does-not-play-diceExperiments on Jev’s probability calibration, uncertainty reporting, and forecast probability preservation.Jev Does Not Play Dice: inspect probabilities on known uncertainty → | Jev Does Not Play Dice: inspect probabilities on known uncertainty | benchmark | JavaScript | 4 | MIT | |
| lirantal/discoprintClassify an artist's discography by theme, mood, and lyrical complexity with Jev (TypeSafe AI), and view it as a colorful terminal dashboardDiscoprint: a classified discography in the terminal → | Discoprint: a classified discography in the terminal | repo | JavaScript | 4 | Apache-2.0 | |
| mpuig/system-oneAn open-source System One decision model — Kahneman's term for the fast, automatic judgment faculty. Calibrated choice/score/noul probabilities in one forward pass on small fine-tuned open models. Local on Apple Silicon (MLX), Jev-compatible API, audited experiment log.system-one: calibrated local models on Apple MLX → | system-one: calibrated local models on Apple MLX | replica | Python | 4 | MIT | |
| tinxai/multimodal-jev-groundingMultimodal Jev prototype: option probabilities plus bounding boxes → | Multimodal Jev prototype: option probabilities plus bounding boxes | replica | Python | 4 | — | |
| zadescoxp/kadeconsoleKade console is a bloomberg terminal type of analytical tool. kadeconsole: a terminal-first stock analytics tool → | kadeconsole: a terminal-first stock analytics tool | product | Python | 4 | — | |
| 0xArx/jevegisGuardrails for LLM apps in one API call. Prompt injection, jailbreaks, leaks, unsafe content. Built on TypeSafe Jev. MIT.Jevegis: guardrails API on Jev → | Jevegis: guardrails API on Jev | product | TypeScript | 3 | MIT | |
| andrewsilber/JevsBistro3D restaurant service simulator for benchmarking low-latency decision modelsJevsBistro: replayable restaurant service decisions → | JevsBistro: replayable restaurant service decisions | demo | TypeScript | 3 | MIT | |
| Anil-matcha/awesome-jev-by-typesafeEvidence-backed use cases, patterns, prompts, and starter code for TypeSafe Jev — a System One model for fast, typed, confidence-aware decisions in software.Awesome Jev by TypeSafe (unofficial) → | Awesome Jev by TypeSafe (unofficial) | directory | Python | 3 | MIT | |
| beamnxw/minelogMineLog harness - GPT-6 Astra plans, Jev decides, one Minecraft body. Backend behind minelog.xyzMineLog: Astra plans, Jev decides, one Minecraft body → | MineLog: Astra plans, Jev decides, one Minecraft body | demo | JavaScript | 3 | MIT | |
| carlaiau/read-with-jevA Demo of using JEV to classify various attributes of a book, and present that to the reader to augment the reading experienceRead with Jev: an emotion-aware novel reader → | Read with Jev: an emotion-aware novel reader | demo | TypeScript | 3 | — | |
| Cloudgeni-ai/genigrepAsk a question about a codebase, get back only the source that answers it. Jev-ranked code search for coding agents.genigrep: source evidence packs for coding agents → | genigrep: source evidence packs for coding agents | integration | TypeScript | 3 | Apache-2.0 | |
| gholtzap/jev-codex-model-and-effort-routerA Jev-based model and effort router for Codex → | A Jev-based model and effort router for Codex | integration | Python | 3 | — | |
| iamvatsalpatel/tiershiftShift every LLM call to the cheapest model that can handle it. Routing decided by TypeSafe Jev in ~180 ms. No training data. Policy in plain YAML. TypeScript and Python.tiershift: policy-bounded LLM cost router → | tiershift: policy-bounded LLM cost router | integration | TypeScript | 3 | MIT | |
| JohGirard/doom-ai-overlordDoom played autonomously by the Laya System-1 decision model (open Jev alternative) via ViZDoomdoom-ai-overlord: Doom on a 421M local decision model → | doom-ai-overlord: Doom on a 421M local decision model | demo | Python | 3 | MIT | |
| legacybridge-tech/pi-typesafe-jevA pi extension that exposes TypeSafe (Jev, System One) judgments as five pi tools, so a model can make narrow semantic judgments while your code and your users keep control of thresholds, weights, and actions.pi-typesafe-jev: Jev as five native Pi tools → | pi-typesafe-jev: Jev as five native Pi tools | integration | TypeScript | 3 | Other | |
| mateonunez/jodSemantic schemas over TypeSafe's Jev — validate the state locally, then project typed answers.jod: semantic schemas over Jev → | jod: semantic schemas over Jev | integration | TypeScript | 3 | MIT | |
| mizchi/jev-lexerLanguage-agnostic syntax highlighter: split like gpu-lexer, classify every part with Jev, render Shiki-compatible tokens, HTML and ANSIjev-lexer: syntax highlighting with no language knowledge → | jev-lexer: syntax highlighting with no language knowledge | demo | TypeScript | 3 | MIT | |
| molis-ai/jev-workbenchBuild versioned judgment functions on TypeSafe's Jev once, then call the same published version from your backend over HTTP and from coding agents over MCP. The vendor key stays on your machine.Jev Workbench: versioned judgment functions → | Jev Workbench: versioned judgment functions | product | TypeScript | 3 | MIT | |
| morinokami/hunk-triageRead the important changes first.hunk-triage: read the important changes first → | hunk-triage: read the important changes first | integration | TypeScript | 3 | MIT | |
| muse0509/jev-preflightA bounded Jev risk check for Claude Code: eight risk axes, one request, one optional reinspection.jev-preflight: a bounded risk check for Claude Code → | jev-preflight: a bounded risk check for Claude Code | integration | Go | 3 | MIT | |
| reycn/smart-switchReimagined window switcher for macOS using frontier artificial intelligence. Predicted by TypeSafe's Jev modelSmartSwitch: predict the next Mac app → | SmartSwitch: predict the next Mac app | integration | Swift | 3 | AGPL-3.0 | |
| TokenTrim/jev-routing-experimentBenchmarking TypeSafe's Jev decision model as a cost-efficient LLM router on RouterArenaJev as an LLM router: a mostly negative result → | Jev as an LLM router: a mostly negative result | benchmark | Python | 3 | Apache-2.0 | |
| TypeSafeAI/clarity-judgeMulti-axis writing quality checker powered by TypeSafe AI's Jev model. Separate named checks, each with its own verdict and confidence.Clarity Judge: seven writing checks → | Clarity Judge: seven writing checks | demo | TypeScript | 3 | — | |
| wdwxw/jevdemojevdemo: Choice, Score and Noul in Python → | jevdemo: Choice, Score and Noul in Python | demo | HTML | 3 | — | |
| AdamPippert/granite-decisionsGraniteJev...granite-decisions: a Granite 4.1 3B RHCD model → | granite-decisions: a Granite 4.1 3B RHCD model | replica | Python | 2 | MIT | |
| altryne/twitter-timeline-analyzerChrome extension to analyze and filter your Twitter/X timeline using AI-powered topic detection with W&B Weave observabilityGot Jev — now what? A real deployment list → | Got Jev — now what? A real deployment list | roundup | JavaScript | 2 | — | |
| FirasSX914/JanusMeasure when to use Jev and other models on your data, then route accordingly.Janus (formerly calibre): confidence-routed fallback study → | Janus (formerly calibre): confidence-routed fallback study | benchmark | Python | 2 | MIT | |
| phyous/tsai-civ2TypeSafe Jev plays original Civilization II in a browser, with live action probabilities. Experimental full-game harness.Jev × Civilization II: an experimental campaign harness → | Jev × Civilization II: an experimental campaign harness | demo | Python | 2 | Other | |
| sawzhang/jev-demoJev (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。sawzhang/jev-demo: five demos and Jev measurements → | sawzhang/jev-demo: five demos and Jev measurements | benchmark | HTML | 2 | — | |
| shunshunNi/redmine_jevAsks Jev whether a Redmine issue's tracker fits, while you fill in the formredmine_jev: does this issue's tracker fit? → | redmine_jev: does this issue's tracker fit? | integration | Ruby | 2 | MIT | |
| southleft/component-charadesA Taboo-style parlour game for design systems, refereed by Jev (TypeSafe System One model)Component Charades: Jev guesses a UI component → | Component Charades: Jev guesses a UI component | demo | TypeScript | 2 | — | |
| suraj-phanindra/wellposedLints Jev requests before they are sent: offline checks for missing none-of-the-above options and broken state paths, then Jev itself for what structure cannot decide.wellposed: lint the structure and meaning of Jev questions → | wellposed: lint the structure and meaning of Jev questions | repo | JavaScript | 2 | MIT | |
| TenzinDhonyoe/bevBev: the slowest classifier in the world, at Jev prices. npx bev-aibev: the slowest classifier in the world, at Jev prices → | bev: the slowest classifier in the world, at Jev prices | demo | TypeScript | 2 | MIT | |
| teyhouse/jev-secret-detectionMeasures how well TypeSafe's RLCD-Jev model spots real secret credentials in file snippetsSecret detection test with Jev → | Secret detection test with Jev | demo | Python | 2 | — | |
| Teylersf/WindowsJevToken-efficient Windows automation and durable research MCP server for Codex and Claude Code, powered by TypeSafe Jev.WindowsJev: bounded Windows actions and durable research → | WindowsJev: bounded Windows actions and durable research | repo | C# | 2 | MIT | |
| TheEleventhAvatar/triage-botReal-time support triage + response bot Jev routes the ticket to a specialist agent (general / account / billing / technical) and decides whether a human should take it instead — all as typed data, no text to parse. Cerebras then drafts the reply using whichever agent Jev picked. The script times both calls separately so you can see the split.triage-bot: real-time support triage with Jev and Cerebras → | triage-bot: real-time support triage with Jev and Cerebras | demo | TypeScript | 2 | — | |
| Tom-R-Main/FootworkBrowser and desktop automation for AI agents, with guarded actions and receipts to check what actually happened.Footwork: Jev guards and mechanical computer-use steps → | Footwork: Jev guards and mechanical computer-use steps | repo | Python | 2 | MIT | |
| wustep/jev-playgroundCan a System One model steer music? Jev picks the plan (enums only); code renders sheet, audio and MIDI.Jev Playground: typed plans rendered as sheet music and MIDI → | Jev Playground: typed plans rendered as sheet music and MIDI | demo | TypeScript | 2 | — | |
| 4esv/jev-evalBenchmark TypeSafe Jev against any OpenRouter model on your own data.Jev vs GPT-5.6 Terra head-to-head → | Jev vs GPT-5.6 Terra head-to-head | benchmark | Python | 1 | — | |
| aadilghani1/qc-useTest the most important path in your app. Describe it in plain words. Powered by TypeSafe Jev.qc-use: Jev-assisted UI tests from plain-language steps → | qc-use: Jev-assisted UI tests from plain-language steps | integration | Python | 1 | MIT | |
| Bhuwan-web/intent-classificationType-safe intent classification guard for agentic workflowsintent-classification: a scope guard in front of agent workflows → | intent-classification: a scope guard in front of agent workflows | integration | HTML | 1 | — | |
| denser-org/rerank-bench-jevProduction reranker benchmark: TypeSafe Jev vs Qwen3-Reranker-0.6B on BEIR SciFact and NFCorpus — nDCG@10, cost/query and p50/p95 latency. Quality is indistinguishable; Jev is ~2x faster, Qwen ~4x cheaper. From the team at denser.airerank-bench-jev: Jev versus Qwen3 Reranker → | rerank-bench-jev: Jev versus Qwen3 Reranker | benchmark | Python | 1 | Apache-2.0 | |
| doom-satellite/DoomSatF´ flight software → CCSDS/Yamcs → Open MCT, with jev (System One) and Claude Sonnet 5 (System Two) driving Doom over that real mission stack.DoomSat: a game payload through a mission-control stack → | DoomSat: a game payload through a mission-control stack | demo | Python | 1 | — | |
| forestwas/gmail-jevGmail inbox triage with TypeSafe Jev — workflow labels, archive decisions, and optional live/backfill workers.gmail-jev: label Gmail threads with Jev → | gmail-jev: label Gmail threads with Jev | integration | Python | 1 | MIT | |
| HackSing/jev-report发明 RLHF 的人，这次做了个不会说话的模型：Jev 独立研究报告。52 页 PDF + 50 条中文实测复现包 + 143 条可回溯数据表Chinese-language fact-check of the 193.6x / 444.6x claims → | Chinese-language fact-check of the 193.6x / 444.6x claims | benchmark | Python | 1 | MIT | |
| Hu-xiao-max/jev_robotjev_robot: a moving-target pick-and-place in MuJoCo → | jev_robot: a moving-target pick-and-place in MuJoCo | demo | Python | 1 | Apache-2.0 | |
| Jabbslad/pi-jev-toolsTypeSafe ranking, classification, retrieval and structured-decision tools for Pi coding agentspi-jev-tools: ranking and retrieval tools for Pi → | pi-jev-tools: ranking and retrieval tools for Pi | integration | TypeScript | 1 | MIT | |
| JacobLinCool/jev-paper-judgeFeedback on your paper in seconds.Jev Paper Judge → | Jev Paper Judge | demo | TypeScript | 1 | MIT | |
| Jevals/jevals-dataIndependent benchmark data for TypeSafe's Jev (System One model) vs LLMs: accuracy, calibration, cost. Boards + per-decision logs, CC-BY-4.0Jevals: labelled benchmarks and per-decision logs → | Jevals: labelled benchmarks and per-decision logs | benchmark | — | 1 | CC-BY-4.0 | |
| lib-x/typesafe-goGo SDK for the TypeSafe AI API: typed questions in, structured answers out. Zero dependencies.typesafe-go: a zero-dependency Go SDK → | typesafe-go: a zero-dependency Go SDK | integration | Go | 1 | MIT | |
| markfive-proto/typesafe-vs-deepseekTypeSafe (Jev) vs DeepSeek-flash: side-by-side speed/token/cost/accuracy comparison across invoice extraction, email classification, and rerankingJev vs DeepSeek-flash side by side → | Jev vs DeepSeek-flash side by side | benchmark | Python | 1 | — | |
| maxlibin/jev-rubiksInteractive 3D Rubik's cube: code solves, TypeSafe's Jev coaches your hand solvesjev-rubiks: the code solves, Jev coaches your hand solve → | jev-rubiks: the code solves, Jev coaches your hand solve | demo | TypeScript | 1 | MIT | |
| oboroge0/jev-voice-games声で遊ぶミニゲーム 2 本 (ボイスランナー / スイカわり) — Godot × hayamimi × Jevjev-voice-games: voice-driven Godot mini-games → | jev-voice-games: voice-driven Godot mini-games | demo | GDScript | 1 | MIT | |
| ppramanik62/yc-jev-benchCan Jev replace an LLM reranker? A benchmark of TypeSafe's Jev vs Claude Haiku and BGE on natural-language search over all 6,245 YC companies, with a live search demo.YC × Jev: startup search and a reranking benchmark → | YC × Jev: startup search and a reranking benchmark | benchmark | TypeScript | 1 | — | |
| thehumanworks/jevgrepjevgrep jg: Rust natural-language code search → | jevgrep jg: Rust natural-language code search | repo | Rust | 1 | — | |
| torumitsutake/jev-page-verdictPage Verdict: classify web pages by purpose and stance → | Page Verdict: classify web pages by purpose and stance | product | JavaScript | 1 | MIT | |
| 0x963D/last-exitA cyberpunk border encounter powered by TypeSafe Jev. Bluff the guard. Inspect the receipts.Last Exit: a checkpoint game judged by Jev → | Last Exit: a checkpoint game judged by Jev | demo | JavaScript | 0 | Other | |
| ajbmachon/jev-navigatorFind code with small, typed Jev judgments: a code index, closed yes/no and pick questions, and a best-first search where code owns the loop.Jev Navigator: code relationships and semantic search → | Jev Navigator: code relationships and semantic search | integration | Python | 0 | MIT | |
| blowxian/jev-fanout-benchMeasured: asking TypeSafe Jev N questions in one call bills the state once. 2,976 real requests, raw data, exact billing check.jev-fanout-bench: measure batching, billing and answer stability → | jev-fanout-bench: measure batching, billing and answer stability | benchmark | Python | 0 | MIT | |
| ChenReuven/xo-jevX/O vs Jev — tic-tac-toe with TypeSafe Choicexo-jev: Jev plays tic-tac-toe → | xo-jev: Jev plays tic-tac-toe | demo | TypeScript | 0 | — | |
| chepyle/jev-testReproducible zero-shot Jev benchmark on all seven LexGLUE tasksJev × LexGLUE: typed decisions across seven legal tasks → | Jev × LexGLUE: typed decisions across seven legal tasks | benchmark | Python | 0 | Apache-2.0 | |
| CorieW/JevTestBounded exploratory browser testing with Jev, deterministic assertions, and replayable evidence.JevTest: exploratory browser tests with replayable evidence → | JevTest: exploratory browser tests with replayable evidence | repo | TypeScript | 0 | — | |
| DomMonte/n8n-nodes-typesafe-ain8n community node for the TypeSafe AI System One API — typed yes/no, choice and score questions with calibrated probabilitiesn8n community node for the System One API → | n8n community node for the System One API | integration | TypeScript | 0 | MIT | |
| enderkus/zammad-jev-dispatcherAI ticket triage for Zammad using Jev (TypeSafe AI) - classifies and routes new tickets into the right service group automatically.Zammad ticket dispatcher on Jev → | Zammad ticket dispatcher on Jev | integration | Ruby | 0 | MIT | |
| gbesse/jev-lab59 open-source projects exploring typed AI decisions with Jevjev-lab: 59 open-source projects around one question → | jev-lab: 59 open-source projects around one question | directory | — | 0 | MIT | |
| gopalanj/chess-jevChessJev: a System One model for chess → | ChessJev: a System One model for chess | replica | Python | 0 | MIT | |
| jagenaujagenau/ground-truthGround News style bias check for the article in your current tab.Ground Truth: news-framing browser extension → | Ground Truth: news-framing browser extension | integration | TypeScript | 0 | MIT | |
| jayanthchandra/jgrepA Rust CLI that brings Jev's probabilistic reasoning to the Unix command line. Pipe logs, files, or any text, ask natural-language questions, and get structured semantic decisions.jgrep: ask questions of logs, files and pages from the shell → | jgrep: ask questions of logs, files and pages from the shell | integration | Rust | 0 | — | |
| jeffclimbsbikes/system-one-playgroundRun the new wave of decision models (Laya, Decider, Kev, Jev) side by side on your Mac. Structured input in, calibrated probabilities out, with honest accuracy, calibration, latency and memory numbers.System One Playground: compare decision engines on Apple Silicon → | System One Playground: compare decision engines on Apple Silicon | repo | TypeScript | 0 | MIT | |
| joaoh82/coffee-under-fireCoffee Under Fire: a free browser shooter with NPC tactics powered by TypeSafe AI’s Jev model. Play at https://coffee.yardsort.sh/Coffee Under Fire: Jev-controlled NPCs in a browser shooter → | Coffee Under Fire: Jev-controlled NPCs in a browser shooter | demo | TypeScript | 0 | MIT | |
| jourdanlabs/assay-001ASSAY-001: independent, pre-registered verification of TypeSafe Jev's calibration and type-safety claims. Split verdict, published in full.ASSAY-001: calibration and type-safety evidence on two intent corpora → | ASSAY-001: calibration and type-safety evidence on two intent corpora | benchmark | Python | 0 | — | |
| juancamiloqhz/roverlabA 3D planetary rover sandbox for experimenting with autonomous decisions using TypeSafe AI.Roverlab: an inspectable planetary expedition sandbox → | Roverlab: an inspectable planetary expedition sandbox | demo | TypeScript | 0 | — | |
| jxspam/jev-ui-mapMap your app's UI so Jev can drive it. Demo: order food by voice. Built with AdaL.jev-ui-map: typed actions for browser tasks → | jev-ui-map: typed actions for browser tasks | demo | JavaScript | 0 | MIT | |
| LoganGeffen/jev-agent-deskLocal Jev-powered conversation and voice interface for Codex and Claude in tmuxJev Agent Desk: local control of Claude and Codex sessions → | Jev Agent Desk: local control of Claude and Codex sessions | repo | Python | 0 | — | |
| lukeramsden/system-oneTyped decisions from System-1 models (Jev, Cloudflare AI Gateway, Laya) with Promise and Effect-native clientssystem-one: Promise and Effect clients for typed decisions → | system-one: Promise and Effect clients for typed decisions | integration | TypeScript | 0 | MIT | |
| m-mizutani/semgateSemantic request filtering and routing for Go HTTP, powered by TypeSafe AI.semgate: semantic request filtering for Go HTTP → | semgate: semantic request filtering for Go HTTP | integration | Go | 0 | Apache-2.0 | |
| marcosmartinez/jev-acento¿Jev entiende tu acento? Pre-registered audit of TypeSafe AI's Jev on Spanish — accuracy, calibration and token cost — plus a CLI to run the same comparison on your own labelled data.jev-acento: a paired Spanish-language accuracy and calibration audit → | jev-acento: a paired Spanish-language accuracy and calibration audit | benchmark | Python | 0 | MIT | |
| mourad-ghafiri/qwen3.5_0.8B_decision_modelA small, fast decision model. You give it a piece of text (the state) and typed questions about it. It answers with calibrated probabilities over the options you define, never generated text, and it takes about 50 ms per request on a GPU.Qwen3.5 0.8B decision model: a local Jev alternative → | Qwen3.5 0.8B decision model: a local Jev alternative | replica | Python | 0 | MIT | |
| pedroknigge/mcp_jevOpen MCP server to run TypeSafe Jev (System One) packs locally — Choice / Noul / Score for Cursor & agentsmcp_jev: local MCP server with Jev packs → | mcp_jev: local MCP server with Jev packs | integration | TypeScript | 0 | MIT | |
| rimusz/jev-mcpProof of concept MCP for Typesafe's new Jev AI modeljev-mcp → | jev-mcp | repo | TypeScript | 0 | MIT | |
| rmosleydb/jev-agent-orchestratorTurn one prompt into a purpose-built agent on Databricks. JEV (TypeSafe System One) picks the model, MCP servers, and skills from Unity Catalog metadata in a single call; the agent is composed with the OpenAI Agents SDK and invoked.Jev Agent Orchestrator: Databricks catalog-based agent routing → | Jev Agent Orchestrator: Databricks catalog-based agent routing | integration | Python | 0 | MIT | |
| RobertKoval/jev-agent-toolsTyped Jev decisions with automatic session context for Pi and Codex. One shared core, compact TOON output.Jev Agent Tools: bounded conversation context for Pi and Codex → | Jev Agent Tools: bounded conversation context for Pi and Codex | integration | TypeScript | 0 | MIT | |
| Ryoseiimai/shibuya-machimachi渋谷駅から半径1.5km限定、相互承認した2人だけで位置を共有する1対1待ち合わせアプリShibuya Machimachi: a Jev-assisted meetup app → | Shibuya Machimachi: a Jev-assisted meetup app | product | JavaScript | 0 | MIT | |
| SkywalkerDarren/feed-lensCustomizable social feed labels for Weibo, Threads and X, powered by your TypeSafe accountFeed Lens: your labels on X, Threads and Weibo → | Feed Lens: your labels on X, Threads and Weibo | product | JavaScript | 0 | Apache-2.0 | |
| TrainLCD/Functions👷 Cloudflare Workers for the TrainLCD mobile app.TrainLCD: Jev re-ranking in a production train app → | TrainLCD: Jev re-ranking in a production train app | integration | TypeScript | 0 | — | |
| Wizhill05/typesafe-image-diffusionDiffusion-style pixel art out of a general classifier (TypeSafe Jev): 256 parallel pixel questions + refinement passesJev as a pixel-art generator → | Jev as a pixel-art generator | demo | HTML | 0 | — | |
| xpressabhi/jev-browserJev decides. The harness acts.xpressabhi Jev Browser: a harness-independent decision core → | xpressabhi Jev Browser: a harness-independent decision core | integration | TypeScript | 0 | MIT | |
| zsoXi/agent-handoff-gate archivedAbandoned project (learning/fun). An experimental protocol for evidence-aware agent handoffs, bounded worker continuation, and TypeSafe/Jev-assisted review, with reproducible evaluation.Agent Handoff Gate → | Agent Handoff Gate | integration | Python | 0 | MIT | |
| choxos/jevchessJevChess: Jev vs LLMs, Stockfish or you → | JevChess: Jev vs LLMs, Stockfish or you | demo | — | — | — | — |
| pymodel/typesafe-mcptypesafe-mcp: Jev as an MCP server for a local GLM agent → | typesafe-mcp: Jev as an MCP server for a local GLM agent | integration | — | — | — | — |
| qiyangdev/hn-jevHN Jev Scores: adjustable Hacker News ranking → | HN Jev Scores: adjustable Hacker News ranking | integration | — | — | — | — |
No repositories match.
Stars belong to the entire repository, including work unrelated to Jev. Totals count each current repository once; indexed addresses retain historical aliases. Check dates vary by repository; view activity and check dates.