Overview
Surveys the model or the space rather than one pattern.
481 rows are filed under this pattern, 459 of them with code. Rows are listed as the catalogue orders them: official first, then rows with code, then by stars.
Filter and search them in the catalogue →
Rows marked “not Jev itself” or “shadow mode” stay listed, but they are not examples of deciding with Jev.
-
Official agent skill for Claude Code
Installs a TypeSafe skill into Claude Code so an agent can write correct Jev calls without you pasting the API shape each time.
-
system-one-adapter-python
A drop-in TypeSafeClient replacement backed by ordinary LLM APIs, so you can run Jev-shaped code without Jev access.
-
typesafe-sdk (Python)
The official Python client. Sync and async clients, retry policy with retry-after support, and Choice/Score/Noul helper classes.
-
@typesafe-ai/sdk (TypeScript / JavaScript)
The official TypeScript client. Ships ESM, CJS and type declarations, with lowercase choice()/score()/noul() helper factories.
-
API reference
POST /v1/systemone request and response shapes for the core Choice, Score and Noul questions, plus request validation and HTTP error handling.
-
Models, pricing and limits
Native Jev 1.14/1.13 IDs, aliases, pricing and limits: $0.042/Mtok input, free output, 64k tokens per request and 32k per question by default. Text is the default; images and 262k long context require experimental opt-ins.
-
Primitives: Choice, Score, Noul
How to choose and combine the core Choice, Score and Noul question types and define their criteria. Choice supports up to 255 options and Score 2-10 levels.
-
Introducing System One models and Jev
The launch post: what a System One model is, why decisions were split from generation, and the vendor's latency and cost claims.
-
Jev 1.13 known limitations
The vendor's Jev 1.13 limitations: literal reading, arithmetic and counting, date comparison, indirection, large noisy states and adversarial content.
-
Use case map
The vendor's own taxonomy: five headline categories, nineteen industry groups, and ten decision shapes from classification through to structured data extraction.
-
OpenCode Zen: Jev resale
A coding agent whose hosted gateway resells Jev, including a free tier model id.
-
litellm
AI gateway that calls 100+ LLM providers through one OpenAI-format API. Its TypeSafe guardrail compacts context: for each finished tool exchange in a request it asks Jev a noul, whether the result is still needed for the task, and blanks the results judged no longer relevant.written by a model
-
laya
A reimplementation of the System One interface, not Jev: its own non-autoregressive checkpoints answer typed choice, score and noul questions over text or JSON in one forward pass, 33 ms for one question on a T4 by the author's measurement, in 100+ languages, with a router that picks the checkpoint.written by a model
-
ai
The AI Toolkit for TypeScript. From the creators of Next.js, the AI SDK is a free open-source library for building AI-powered applications and agentsupstream description
-
openwork
The open-source alternative to Claude Cowork (powered by opencode)upstream description
-
@effect/ai-typesafe
Implements Effect's DecisionModel interface over Jev, with an unusually candid caveat about unverified rounding behaviour.
-
kev
A trainable, self-hostable family of Jev-like decision models with a System One compatible API, so the official SDK can point at your own server.
-
rig-typesafeai
A Rust integration with compile-time-checked option counts, so an over-255 Choice fails to build rather than at runtime.
-
deep-searcher
Deep-research tool that searches private data in a vector database such as Milvus and reasons over it with an LLM. Its evaluation compares the LLM's own decision to stop searching with a Jev noul on whether the gathered context is enough to answer, over 100 questions.written by a model
-
Yao Agents
Self-hosted agent workspace that runs agents across your own devices with a task board, isolated workspaces and an open API; it added Jev as a decision-model provider alongside its text models. The cited file tests Jev key validation against /v1/models.written by a model
-
Kiln: Jev adapter
A JSON-Schema-to-question compiler wired into the adapter registry, with an honest note on what it cannot serve.
-
SemIf
An independent semantic-if implementation that states up front it is unaffiliated with Jev or TypeSafe.
-
ruby_llm: TypeSafe provider
A Ruby provider with a dedicated System One protocol, the main route into Jev from Ruby.
-
NanoJev
A self-described nano replica of Jev, for reading rather than for production.
-
NeoHorse
TokenRhythm's open-weight model family for agent workflows. NeoHorse-Jev-4B, built on NeoHorse-1-4B, answers Choice, Noul and Score by prefill-only inference behind a /v1/systemone-shaped runtime for vLLM, SGLang or Python. The authors report 77.70 across six text benchmark groups.written by a model
-
Jeff
An open 0.8B 'System 1' model with task adapters and GGUF builds, meant to sit in front of a large local model and make quick decisions; its game runner asks one choice per turn over legal moves via /v1/systemone. The author reports 94.6% mean accuracy over 8 adapters vs 86.6% for Qwen3.8-27B.written by a model
-
Ollaya
Runs open decision models locally the way Ollama runs LLMs: pull a model by name and serve it behind a /v1/systemone-compatible daemon, so the TypeSafe SDK works unchanged. The README reports 0.724 typed-decision accuracy for its recommended winnow:e4b against TypeSafe's 0.738 for Jev.written by a model
-
jevlike
An independent, trainable model with the same input and output shape as Jev: text plus N options in, one probability per option out, in a single pass.
-
decider
A family of System One-style models fine-tuned from an open base for one-pass typed decisions.
-
rizzo-flow
Local reimplementation of Jev's interface, not Jev: the open Spark-X2.5 model with the author's LoRA fine-tune, on llama.cpp, answers noul, choice and score questions on a Jev-style endpoint. On typed-decisions it scored 0.648 to the base model's 0.574, below Jev; probabilities are uncalibrated.written by a model
-
jev (jevos)
One CPU-only binary that serves an open model, jevos-v5, behind Jev's POST /v1/systemone: noul, choice and score answers with probabilities in about 30–130 ms on a laptop, no Python. Weights on Hugging Face and as GGUF; the author's 7-task benchmark has Jev ahead on accuracy.written by a model
-
von
Archived reimplementation of the System One interface, not Jev: a 395M-parameter ModernBERT encoder that answers choice, noul and score questions in one forward pass on CPU or GPU, served on a /v1/systemone endpoint, so code written for Jev can switch by changing the base URL.written by a model
-
localjev
GitHub Next's local, Jev-compatible POST /v1/systemone API in TypeScript, reading probabilities from a DiffusionGemma model through a one-step structured read on a patched vLLM. It reimplements the wire, not the model.
-
vLLM Jev
vLLM-based serving engine for open Jev-shaped decision models such as Valen: scores choice, noul and score candidates through prefill with shared-prefix reuse, with image and video inputs and an MLX path for Apple Silicon. The author reports 4.3× over Valen's PyTorch reference on one video run.written by a model
-
openjev
A Jev-compatible decision server on an open diffusion model.
-
Strands Decider
Open decision model for Strands Agents workflows: a 2B Qwen3.5 model, plus Gemma 4 sizes up to 26B-A4B, served behind a /v1/systemone-shaped endpoint and answering noul, choice and score with a confidence. The authors report 180 of 231 public JevBench tasks for the 2B and 207 for the 26B-A4B.written by a model
-
simple-jev
Turns any open-weights model into a Jev-shaped endpoint by reading next-token logits, with the server constructing the JSON rather than the model generating it.
-
Open-Jev
Open-Jev-27B, an open-weight model that returns typed probabilities for a context, questions and candidates without generating an answer — a Jev-style decision model you can run yourself.
-
Jeeves
PostHog's open 9B Jev-like model (Qwen3.5-9B, LoRA, pointer head) that answers noul, choice and score directly when confident and reasons first when not, with a diffusion drafter, training code and data. The authors report it ahead of Jev on JevBench's public tiers.written by a model
-
imajev
Small open models (2B, 4B, 9B) that read the photos, records and text a business already has and answer in the options you set, each with a probability and an explicit can't-tell, so a system acts when sure and hands the rest to a person. The author reports first place of 91 on JevBench v1.4.2.2.written by a model
-
Phocinae-Largha-150M-v1
A 144.3M-parameter bilingual (English and Chinese) typed decision model fine-tuned from mmBERT-small, answering noul, choice and score with zero output tokens through a /v1/systemone-compatible server, and runnable on CPU. The card leads with 0.791 typed accuracy in English and 0.848 in Chinese.written by a model
-
openjev-sglang
A Jev-compatible endpoint served from open models, prefill only.
-
jevbench
Benchmark Heaven's benchmark for Jev and Jev-compatible decision models: noul, choice and score tasks, sent through a /v1/systemone adapter and scored for accuracy, calibration, latency and cost. Release v1.6.1 (2026-10-06); some inputs are sealed, so it cannot be fully rerun independently.written by a model
-
jeff
Self-hosted reimplementation of Jev's System One API, not Jev: GLiFormer, a 400M-parameter encoder, answers choice, score and noul questions, so the official SDK works by pointing its base URL at it. Cheaper but less accurate: 75.5% on AG News topics against Jev's 90.5%, by the author's run.written by a model
-
10 Levels of Jev
Thirty uses of Jev arranged in ten levels, from one smart if statement in plain code to hooks and tools inside the pi coding agent and finally an agent that writes its own questions. A live lab, terminal runner, offline mock with 201 tests and a companion YouTube video.written by a model
-
jiwo
Open decision models on Qwen3.5 (jiwo-0.8b and jiwo-4b) that answer Jev's POST /v1/systemone request with a probability for every option in one forward pass, generating no text. The author reports public Decision Index 0.3 scores of 29.58 and 46.06 and a median request of about 50 ms on one H100.written by a model
-
OneJev
Multimodal open decision model on Qwen (0.8B, 4B, 9B and 27B) that returns calibrated probabilities for typed questions about screenshots, photos, video and text, with a local server the official TypeSafe SDK talks to unchanged. The authors report 69.49 for the 27B on the public Decision Index 0.3.written by a model
-
OpenJev
Open-source playground for System One-style questions, not Jev: you write a state and choice, score and noul questions and get typed answers from one of three backends, an LLM scoring each option, one JSON call to any OpenAI-compatible model, or Mapika's open decider weights.written by a model
-
jevk5
Open-weight reimplementation of Jev's interface, not Jev: Qwen3.5-4B with a distilled LoRA answers noul, choice and score questions through a TypeSafe-style endpoint, about 13 ms on an H100. JevBench v1.4 ranked its v0.2 second of 76 systems, at 62.04 to Jev's 63.29.written by a model
-
open-jev
Local reimplementation of System One-style option scoring, not Jev: Gemma 3 4B prefills the context once and scores every pre-written option in one forward pass, a softmax giving each option's probability, on MLX or PyTorch, with per-task fine-tuning. A Doom demo picks actions this way.written by a model
-
RSI-Jev
An agent loop that proposes hypotheses, pre-registers predictions, trains and retires open decision models answering noul, choice and score in the /v1/systemone shape; 496 experiments written up, failures included. The author reports 65.64 on the public Decision Index 0.3 set for the 27B VL model.written by a model
-
JevAny
JevAny: a calibrated decision layer for RL, agents and model harnesses that returns typed answers and option probabilities in one prefill pass — an open 27B model on a Qwen backbone, not Jev.
-
djev
Not Jev: a server answering Jev-style questions with DiffusionGemma on vLLM. The answer template is seeded on the canvas with only its slots left as noise, so one denoise step gives a distribution per slot. It serves Jev's request and answer shapes for noul, choice and score, plus span questions.written by a model
-
OpenDecision
An open-source semantic decision engine running a local zero-shot model, with a FastAPI server proven wire-compatible with the official SDK.
-
laya-server
Not Jev: a self-hosted server and web console for Laya's decision models that takes Jev's request format at the same path, with API keys, usage stats and the multilingual model in one Docker image. Answers come from Laya, so thresholds tuned on Jev's noul, choice or score answers may not carry over.written by a model
-
jevcore
Jev for DeepSeek Harness, other MCP hosts and plain Node: a core library of noul, choice and score calls, and a DSH plugin and MCP server exposing the same three Jev tools with two opt-in gates. Nothing is sent until a live provider is set up, and the startup log says what will be sent where.written by a model
-
advocaat
A small typed client for asking questions about your own data.
-
system-one
Provider-neutral System One runtime for TypeScript and Piupstream description
-
JevEmbed
Toolkit that turns embedding models into choice, score and noul answerers: data synthesis, fine-tuning, a /v1/systemone server, a benchmark module and a playground. Released fine-tuned Qwen3-Embedding and KaLM checkpoints and a 1.67M-question dataset on Hugging Face.written by a model
-
Lichen
Drop-in, API-compatible replacement for Jev that runs open-weight models locally behind the same /v1/systemone endpoint, images allowed in the state. The author reports 204 to 207 of 231 public JevBench decisions with gemma-4-26B-A4B on a 24 GB laptop GPU, against 200 for Jev 1.13.written by a model
-
ruby_decision_model
Ruby client for decision models such as Typesafe Jevupstream description
-
sys1
System One compatible API for open decision models, e.g. Laya, written in Rust.
-
jevify
Supersimple way to serve LLMs as a Jev-like endpointupstream description
-
Drex
Open-weight decision model from Nace.AI, a Qwen3.6-35B-A3B MoE with a Kev pointer head, that answers choice, score and noul questions in the System One request format over states of at most 131,072 tokens; self-hosted through Nace's llama.cpp or Ollama forks, or through Nace's hosted API.
-
litjev
Turn any off-the-shelf LLM into a Jev -like decision layerupstream description
-
OpenSourceJev
Turning an LLM model into a Jev like System.upstream description
-
go-jev
Go SDK and CLI for TypeSafe Jev: typed decisions (yes/no, choice, score) from a modelupstream description
-
solar-mini4-jev
A drop-in wrapper that exposes Upstage's Solar models through the Jev System One API shape — noul, choice and score with the same schema — with a hosted bring-your-own-key endpoint.
-
OpenJev
OpenJev: Open-source alternative to TypeSafe Jev. Typed probabilistic decision API (Choice, Noul, Score) powered by open LLMs & constrained logprob calibration.
-
jev_local
Replicating Jev with a local LLMupstream description
-
typesafe-ai-benchmark
A gateway that mimics the structured-output shape, used to benchmark against it.
-
jev (Elixir/OTP)
Jev as an OTP process: reply from a GenServer and pattern match on the answer.
-
jev-explained
Jev Explainedupstream description
-
refgarden
A spatial reference explorer for creators. Local Jev query choices, metadata highlights and source-linked collections.upstream description
-
pijev
Averages Jev's answers over option orderings — all permutations in one request — with Brier score and log loss guaranteed no worse than the average across the orderings included. A one-line import change.
-
Jev-Quantum
A Jev-protocol random baseline: it speaks noul, choice and score but reads no prompt, answering from a fast pseudo-random generator — a mock, and a lower bound for routing evaluations.
-
tinyjev
A tiny jev-like model that answers Choice, Score and Noul questions in one forward pass and returns calibrated probabilities. MLX or PyTorch, fully offline, System One compatible.upstream description
-
typesafe
TypeSafe SDK in Swiftupstream description
-
OpenJev
Jev-compatible System 开源Jev
-
notjev
Super fast Jev like server, model agnostic, working with any OpenAI compatible endpointearlier upstream description
-
fastjev
SDK-first, independently maintained SemIf fork for fast, self-hosted semantic decisions.upstream description
-
Jev-Style
Small local decision models on Qwen3.5 (2B and 0.8B) with a /v1/systemone server, an MCP server with decide, noul, choice and score tools, a Claude Code PreToolUse guard and an eval command for your own labels. The author reports 73.6% on JevBench v1.4.1's public items, self-run.written by a model
-
learn-jev-end-to-end
Learn Jev end to end: a free hands-on course. Build 13 AI agent use cases with a fast brain (Jev) and a slow brain (LLM). One OpenRouter key.upstream description
-
hunch
Probabilistic control flow for Ruby and Rails - powered by TypeSafe's Jevupstream description
-
jev-cli
Small dependency-free CLI for TypeSafe Jevupstream description
-
jevper
Jev-shaped (TypeSafe System One) classification wrapper over OpenAI-like clients: probabilities and confidence instead of prose
-
ruby_llm-typesafe
A structured-output provider for a Ruby LLM library.
-
swift-typesafe
Unofficial Swift SDK for TypeSafeupstream description
-
jev-cookbook
The complete cookbook for Jev by TypeSafe AI — 120+ use cases, 10 runnable examples, 4 composition patterns, and first-principles theory for the world's first System One AI model.upstream description
-
swift-jev
A Swift client for TypeSafe AI's Jev — typed judgements, not textupstream description
-
typesafe-ai
Typed TypeSafe AI clients for Rust, with async and blocking backends and observable retries.upstream description
-
snapjudge
Typed decisions (choice / score / yes-no) from local Qwen models on Apple Silicon. Probabilities come straight from the logits, no text generation. TypeSafe-compatible HTTP API, runs on MLX.upstream description
-
WaterSheep
Open, Apache-2.0 alternative to Jev: a ModernBERT fine-tune served locally on POST /v1/systemone that answers noul, choice and score questions, plus multi-label ones, with a probability for every option.
-
midscene-jev-runner
Community-maintained JEV runner integration for Midscene Testupstream description
-
ruling
Typed, calibrated decisions from a local model. No text generated.upstream description
-
typesafe-sdk-go
unofficial go sdk for typesafe ai. not affiliated with or endorsed by typesafe ai. a side project built to fill the missing go sdk gap, for the community to use.upstream description
-
jev4k
A Kotlin DSL and client for TypeSafe's Jev modelearlier upstream description
-
typesafe-sdk-java
Community Java client for the TypeSafe System One API (unofficial)
-
typesafeai-dotnet-sdk
Community .NET SDK for the TypeSafe AI System One API — typed noul, choice, and score questions with structured, confidence-scored answers. Not affiliated with TypeSafe AI.upstream description
-
Decis
Self-hostable inference server for open decision models that implements the /v1/systemone contract, so the official TypeSafe SDK only needs base_url changed. One Docker image per engine with weights inside (laya-multilingual, kev-0.8b and two Jeff builds), plus a game playground.written by a model
-
diffusion-jev-sglang
A Jev-like decision engine powered by DiffusionGemma and SGLang. Jev with eyes.upstream description
-
dohnuts.cpp
The same decisions, on CPU. System One model that can run on your Personal Computer.upstream description
-
legalforecastbench
LegalForecast-MTD benchmark alpha and official evaluation workflowsupstream description
-
typesafe-ai-php
PHP SDK for Jev: a TypeSafeClient, a fluent SystemOneRequest builder for state and typed questions, and a SystemOneResult keyed by question ID. PHP 8.2+, installed with Composer.written by a model
-
awesome-jev-prompts
Question-design pattern library for Jev typed decisions: 51 entries / 9 categories (choice, score, noul, routing, guardrails, calibration, embodied & multimodal, anti-patterns), bilingual EN/中文, each entry sourced under a four-level verification policy.
-
go-system-one
when a gopher met Jevupstream description
-
jev-android
A Kotlin Android SDK for UI automation powered by TypeSafe Jev, with an accessibility runtime and sample app.upstream description
-
kojev
Kotlin Multiplatform client for Jev that returns your own enum/sealed types instead of string keys.upstream description
-
typesafe-sdk-go
Go SDK for the TypeSafe AI API — typed questions in, probability distributions out.upstream description
-
jev-benchmark
Benchmarks and a playground for TypeSafe's Jev (System One) model: chess, and who-is-the-player-talking-to for speech-to-text game NPCsupstream description
-
jevgo
Unofficial Go client for the TypeSafe AI System One API (Jev) — typed questions in, calibrated answers out.upstream description
-
snap
Typed decisions from unstructured state: one forward pass, zero generated text. Local, deterministic, Jev-compatible. Not affiliated with typesafe.ai.upstream description
-
typesafe-client (Go and Rust)
Unofficial Go (stdlib only) and Rust clients for the System One API with typed noul, choice and score questions, the official SDKs' retry policy, typed errors with request IDs, response validation, GET /v1/models and gateway support.written by a model
-
typesafe-ai-rs
Independent async and blocking Rust SDK for the TypeSafe AI System One APIupstream description
-
typesafe-sdk
Ruby client for typesafe.aiupstream description
-
jev-java
Idiomatic Java SDK for TypeSafe AI Jev System One decision engineupstream description
-
jev-java
Unofficial Java SDK for TypeSafe Jev and Vercel AI Gateway, with Spring Boot and WebClient supportupstream description
-
jev-korean-benchmark
Reproducible early-access evaluation of Jev on Korean understanding and medical text, with runtime and cost evidenceupstream description
-
jev-little-airways
A show-and-tell capability study for Jev, TypeSafe's System One decision model.upstream description
-
jev-sdk-java
Type-safe Java 21 client for the TypeSafe AI Jev (System One) decision APIupstream description
-
sysone-bench
First independent head-to-head benchmark of System One decision models (Laya vs Jev) on byte-identical inputsupstream description
-
typesafeai.net
.NET SDK for the TypeSafe AI platformupstream description
-
cu-Jev
cuda-Jev — a CUDA-native Jev System One decision inference engine. Jev compatible API, examples, and reproducible benchmarks.upstream description
-
decision-bench
Open benchmark runtime for document-grounded decision modelsupstream description
-
jev
A Go client for the TypeSafe AI's System One API and its model, Jev.upstream description
-
jev-go
Go client for TypeSafe's System One API and its model Jev: typed judgments and calibrated probabilities instead of generated textupstream description
-
jevlang
The simplest way to write decision workflows in Python. Python with a smart if.upstream description
-
OpenJev
OpenJev: an independent Jev-inspired System One decision API based on TypeSafe.ai concepts. Choice, score and noul primitives, local mock server, Python and TypeScript SDKs. Real inference planned; not affiliated with TypeSafe AI.upstream description
-
typesafe-ai-go
Community-maintained Go SDK for the TypeSafe AI System One evaluation API, with typed questions, fluent builders, retries, and examples.upstream description
-
typesafe-sdk-swift
Unofficial Swift library for the TypeSafe APIupstream description
-
typesafe_sdk (Elixir)
An Elixir port of the official SDK.
-
werr
Zero-memory System-1 decision engine & TypeSafe Jev wire-compatible runtime powered by Mandelbrot wave dynamics (The Zero-VRAM Gauntlet).earlier upstream description
-
jev-agent-failure-benchmark
Benchmarking Jev (Typesafe.ai) against a strong LLM on the Who&When Pro agent-failure-attribution benchmark (text subset).upstream description
-
jev-does-not-play-dice
Experiments on Jev’s probability calibration, uncertainty reporting, and forecast probability preservation.upstream description
-
jev-rs
Async-first Rust SDK for the TypeSafe AI APIupstream description
-
origin-civilization
AI life-and-civilization simulation: TypeSafe Jev makes every decision (typed, probabilistic, auditable); LLMs plan — OpenAI-compatible APIs, local models (Ollama, LM Studio), Claude Code, Codex.upstream description
-
SystemOneDotNet
.NET client for TypeSafe System One (Jev) — typed questions in, typed answers with probabilities and confidence out. No prompt engineering, no output parsing.upstream description
-
typesafe-ai-java
Community-maintained Java SDK for the TypeSafe AI System One (Jev) API. Not an official TypeSafe product.upstream description
-
typesafe-go
Idiomatic Go SDK for the TypeSafe AI APIupstream description
-
typesafe-sdk-swift
Swift SDK for TypeSafe AIupstream description
-
jev-go-sdk
Dependency-free Go client for TypeSafe AI's System One API and the Jev modelupstream description
-
jev-research
Practical guide to using TypeSafe Jev with Herdr and Claude, Codex, Hermes, and browser agents.upstream description
-
jev4mellea
Jev adapter for Melleaupstream description
-
jevgo
Go client for TypeSafe AI's System One API (Jev), with optional Langfuse instrumentationupstream description
-
jevsbistro
3D restaurant service simulator for benchmarking low-latency decision modelsupstream description
-
typesafe-ai-rails
Community Rails integration for TypeSafe AI's System One API on the community typesafe-sdk gem: Rails configuration, persisted usage and cost telemetry, and an opt-in confidence policy for Choice and Score answers.
-
typesafe-sdk-dotnet
Unofficial .NET port of the TypeSafe AI client SDK (typed questions & answers)upstream description
-
typesafe-sdk-rust
Rust SDK for the TypeSafe AI APIupstream description
-
audio-jevlike
Prosodia: an audio-native Jev-shaped decision model — typed calibrated decisions from speech, no ASRupstream description
-
chat2jev
Convert legacy chat completion API request to Typesafe jev APIupstream description
-
hunch-js
Jev judgments as TypeScript functions over arrays: classify, score, check, where, extract, pick, rank, verify. LLMs propose, Jev decides.upstream description
-
jear
Jev-routed client for NEAR AI Cloud inference and IronClaw agents.upstream description
-
jev-php-sdk
PHP SDK for TypeSafe's Jev: send text and typed questions, get typed answers with calibrated confidence. PHP 8.1+, works with any PSR-18 client, Laravel 8–13.upstream description
-
jev-routing-experiment
Benchmarking TypeSafe's Jev decision model as a cost-efficient LLM router on RouterArenaupstream description
-
jevclient
Async Python client for TypeSafe Jev. Typed questions in, probabilities and choices out, no prose to parse.upstream description
-
s1_ruby
Makes S1-model 'measurement' (and the collapse that follows it) a Ruby primitive.upstream description
-
typesafe
An idiomatic Elixir client for the TypeSafe AI APIupstream description
-
typesafe-sdk
PHP & Laravel SDK for TypeSafe AI's JEV Model seriesupstream description
-
typesafe-sdk
An unofficial .NET SDK for TypeSafe AI, published on NuGet, with deterministic question building and high-throughput verification.
-
antigravity-mcp-semantic-search-with-typesafeai
Fast semantic code search & diff sanity auditor for AI coding assistants (Antigravity, Cursor, Claude Code) powered by TypeSafe System One.upstream description
-
can-jev-bayes
Jev Bayes, No? Testing TypeSafe AI's Jev against Bayesian-optimal strategies, and testing if Jev can effectivly use Bayesian priors.upstream description
-
decido
Probabilistic decisions for Python. Use Jev or bring your own provider; crawl with Playwright.upstream description
-
decision-circuits
Decision circuits: typed questions to a System One model, calibrated probabilities back, gates in code. Zero-dependency Python SDK with LangChain, OpenAI Agents, and Claude Agent SDK integrations.upstream description
-
jev
Unofficial Go client for TypeSafe's System One API and its model, Jev.upstream description
-
jev-go
Unofficial Go SDK for TypeSafe AI's Jev APIupstream description
-
jev-jp-address
Jev (TypeSafe) 性能評価プロジェクト — 日本郵便 KEN_ALL をマスタに、AI SDK 経由の Jev が住所のあいまい一致にどこまで使えるかを検証upstream description
-
jev-lab
Experiments on TypeSafe Jev (System One decision model) via OpenRouter: repeatability, perturbation, and LLM baseline comparisonupstream description
-
jev-lab
Hands-on research lab for TypeSafe's Jev (System One model): reproducible benchmarks of Noul/Choice/Score primitives, confidence gating, fan-out latency, agent control — plus a living audit of the Jev ecosystem.upstream description
-
jev-sim
Jev-compatible /v1/systemone server reading typed decisions from LLM logits, benchmarked against TypeSafe's Jev on the same items via JevBenchupstream description
-
jev-symfony-bundle
Unofficial Symfony bundle for TypeSafe AI's Jev: typed client, validator constraints, Messenger, Workflow guards and profiler panelupstream description
-
kunobi-jev
Rust client for the TypeSafe System One API (Jev)upstream description
-
typesafe-ai-ruby
Ruby client for the TypeSafe AI(Jev) System One APIupstream description
-
typesafe-client
Unofficial typed async Rust client for the TypeSafe System One APIupstream description
-
typesafe-go
unofficial go SDK for typesafe AI, with typed answers, retries, and context supportupstream description
-
typesafe-go
A zero-dependency community Go SDK, including a static analyser that flags poorly designed questions at compile time.
-
typesafe-rs
Latency-first Rust SDK for TypeSafe System One.upstream description
-
typesafe-sdk-go
Go SDK for TypeSafe AI.upstream description
-
typesafe-sdk-go
TypeSafe.AI Go SDKupstream description
-
typesafe-sdk-ruby
Unofficial Ruby SDK for the TypeSafe AI API (Jev model) - typed questions, retries, and typed errors. Community port of typesafe-sdk-js.upstream description
-
typesafe-sdk-swift
typesafe-sdk-js and typesafe-sdk-python port for swiftupstream description
-
typesafe_sdk_ex
Typesafe AI SDK in Elixir using Requpstream description
-
typesafeai-go
Go SDK for TypeSafe AI API https://docs.typesafe.ai/apiupstream description
-
what-is-jev
Independent, source-linked research on TypeSafe AI's Jev (System One), with 947 rubric-scored public repositories, recurring patterns, datasets, and bilingual documentation.upstream description
-
@ai-sdk/typesafe-ai provider
The AI SDK provider package for calling TypeSafe directly, with a sample covering all three question types and nested criteria shapes.
-
awesome-ai-quant
Curated reading map for AI engineers entering quantitative finance: trading agents, LLM factor mining, decision models in markets, tabular FMs — each entry annotated with why AI engineers should care.
-
Bud Decision Studio
Desktop app for macOS, Windows and Linux that installs a local engine and runs eleven open decision models (Kev, Laya, Lev and others) behind TypeSafe's /v1/systemone API and the OpenRouter and Vercel gateway routes, with a playground, evaluation, decision history and a trainer.written by a model
-
Build Your Own JEV Locally: Run a 100% Private AI Agent on Your Machine
Despite the title, this does not run Jev. It builds a Jev-like decision engine from an open LLM using constrained next-token scoring.
-
Jev Explained: How to Add Fast, Typed Decisions to an AI Agent
A third-party explainer with a useful architecture sketch and an unusually honest list of cases where you should not use a decision model.
-
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.upstream description
-
jev-bun1
TypeSafe の Jev を TypeScript SDK で使ってみる最初の 1 歩upstream description
-
jev-cyrillic-audit
Does TypeSafe's Jev keep its accuracy and calibration on Russian? Independent RU vs EN audit (ECE, reliability diagrams, paired bootstrap) on parallel human-labelled data.upstream description
-
jev-lab
TypeScript experiments, evaluations, and latency benchmarks for TypeSafe's Jev modelupstream description
-
jev-no-enem
Reproducible benchmark evaluating TypeSafe AI's Jev (System One paradigm) on Brazil's ENEM 2025 standardized exam. Evaluates typed decision-making, domain-specific accuracy, and RLCD uncertainty calibration against open LLM baselines with an interactive GitHub Pages dashboard.upstream description
-
Jeva
Research System One model from the National University of Singapore (Jeva-4B). The encoder reads the state and each option separately, so option order does not matter, options can be cached and a null answer is explicit. Its Python `system_one` takes Jev-style Choice, Noul and Score questions.written by a model
-
laya-php
PHP 8.4 and Laravel SDK for Laya, the open multilingual decision engine: describe a decision as a readonly class with enums, and laya-serve's /v1/systemone endpoint answers it as a typed object with a confidence per answer. It adds a fake for tests, caching, events and batch prediction.written by a model
-
OpenJev (alanhuangyoo)
Open decision models (OpenJev-1.7B, 4B, 8B; package wev-ai) behind POST /v1/systemone, distilled from a browser agent's typed requests on live sites. Besides choice, yes/no and score, they pick a browser agent's next operation and element. The author reports 39 to 41 of 153 live tasks for the 4B.written by a model
-
StartLux-Decision
StartLux's open decision models (0.8B to 27B) with an engine and HTTP server answering choice, noul and score behind a /v1/systemone-compatible API on vLLM, llama.cpp or MLX. AutoThink reasons only below a probability gate. The authors report ranking first on Decision Index 0.3.2.written by a model
-
system-one-adapter-rust
Rust port of TypeSafe system-one-adapter (LLM-backed system_one evaluations)upstream description
-
tinyjevclient
An integration with jev by typesafe.ai in Rust
-
Tracing Jev calls with Langfuse
The only platform with dedicated Jev observability: an OpenInference instrumentor that traces every decision call over OpenTelemetry.
-
TypeSafe AI Jev now available on AI Gateway
Vercel's launch note for Jev on AI Gateway, with an experimental_evaluate sample using the model string typesafe-ai/jev.
-
TypeSafe models in Pydantic AI
First-party Pydantic AI support: an Agent with output_type=bool over the typesafe:jev-latest model string.
-
TypeSafe pass-through on LiteLLM
Proxy Jev through LiteLLM for unified keys and cost tracking, with any path under /typesafe/ passed straight through.
-
TypeSafe-compatible API on Vercel AI Gateway
Point the official TypeSafe SDK at Vercel by changing one baseURL, or call the gateway's systemone endpoint directly with cURL.
-
typesafe-go
Unofficial Go SDK for TypeSafe AI's System One API (Jev).upstream description
-
typesafe-sdk-go
Unofficial Go SDK for the TypeSafe AI System One API — 1:1 parity with the official JS and Python SDKs. Not affiliated with TypeSafe AI.upstream description
-
typesafe-sdk-kotlin
A Kotlin SDK for TypeSafe AI, ported from the official JavaScript SDK with its deviations documented; ask typed questions about text and get typed answers back.
-
typesafe-sdk-php
Unofficial PHP library for the TypeSafe APIupstream description
-
typesafe-sdk-php
Unofficial PHP SDK for the TypeSafe AI System One API — 1:1 parity with the official JS and Python SDKs. Not affiliated with TypeSafe AI.upstream description
-
typesafe_ai
An Elixir client for TypeSafe AI with typed responses and bounded concurrencyupstream description
-
typesafe_ai
Typed Elixir client for TypeSafe AI and its Jev System One model, with offline test stubs, concurrent fan-out, and atom-keyed answers.upstream description
-
awesome-jev (yibie)
Currently the most-starred sibling directory in this space.
-
awesome-jev (heyjunpenn)
The broadest sibling directory: hundreds of projects in six languages, with a README that is itself the parsed data source.
-
awesome-jev-projects
A sibling directory aiming at ecosystem breadth with commit-pinned sources, four README languages and a generated site.
-
A new kind of AI model from a ChatGPT inventor is thrilling developers
The only launch coverage with first-hand developer quotes rather than vendor figures, including a caution that interpreting the thresholds is now your job.
-
AI model "Jev" to make machines decide faster
Focuses on the missing explainability — the model returns no reasoning in language — and on every published benchmark coming from the vendor.
-
Decision Models
Hosted API that takes Jev's POST /v1/systemone request shape and answers noul, choice and score questions with probabilities from open models (s1-fast, s1-pro, quyet-1-large); EU tier available, from $0.019 per 1M input tokens, output free.
-
Hacker News: Introducing System One Models and Jev
The launch thread, and the densest single collection of scepticism: unsupported RLCD claims, apples-to-oranges latency comparisons, and the deliberate absence of public benchmarks.
-
Jev (AI model) on Wikipedia
Most useful as an index: its reference list is a fast route to the coverage worth reading.
-
Jev by TypeSafe: A Decision Model for AI Agents
An agent-builder's framing of where a decision model sits in an agent stack.
-
Jev Cuts AI Decision Costs 100x And Vercel, Cloudflare Rushed To Add It
Mainstream coverage of the launch and the speed with which gateways added support.
-
Jev From TypeSafe is a New Class of AI Model that is FAST and CHEAP - But There is a Caveat!
A review that puts the limitation in the title rather than burying it.
-
Jev in the Wild: A Data-Driven Analysis of the Jev Model's Functionality, Applications and Ecosystem
The first data-driven survey and analysis of Jev's application ecosystem examines 2,170 public GitHub projects, documenting rapid early growth, application domains, and decision-use patterns.
-
Jev: System One models for Prod, not God
The only long-form founder interview: why RLHF was the wrong optimisation target, why public benchmarks were withheld, and the all-synthetic data approach.
-
Jev: TypeSafe's System One Model Explained
A neutral survey of the architecture, the claimed benchmarks and the pricing, which states plainly that no large independent reproduction had surfaced.
-
jevai.org community app gallery
Thirty-six community builds curated from social posts: browser agents, spreadsheet tooling, inbox search by intent, ad blocking with judgement, games and robotics.
-
jevai.org community site
An unaffiliated community site with a playground, a preset decision API, an MCP server, downloadable skills and a gallery of community apps.
-
RLCD explained: Reinforcement Learning for Calibrated Decisions
An independent write-up whose most useful finding is a negative one: there is no paper, no reward function, no dataset description and no reproducible evaluation for RLCD.
-
TypeSafe AI debuts model for machines that plays Doom
The most sceptical mainstream piece: it challenges the no-hallucination framing on the grounds that a well-formed answer is not the same as a correct one.
-
TypeSafe on OpenRouter
OpenRouter's listing for Jev, with its own model ids and the unusual pricing shape of paid input and free output.
Not yet indexed by pattern
Projects and plugins with code that nobody has yet read against the patterns: overview is also what the keyword rules suggest when nothing matches.
-
typesafe-ai/skills
The official agent-skills repository behind the Claude Code plugin, holding the SKILL.md that teaches an agent the System One API.
-
langchain
Framework for building agents and LLM apps. Its TypeSafe package wraps Jev as a LangChain runnable: one request asks noul, choice and score questions about text, JSON or chat messages and returns each answer's probabilities and confidence, so code can act, route or ask a person.written by a model
-
oh-my-pi
⌥ Coding agent with the IDE wired in
-
Opik TypeSafe tracker
Wraps the sync and async clients so every system_one call is recorded as a traced span.
-
pydantic-ai
Python agent framework, typed end to end, where any model is a string swap away. Its TypeSafe model runs Jev as a decision model: an agent whose job is to decide something uses Jev like any other model, with its output type and tools turned into choice, noul and score questions.written by a model
-
eliza
Open source agentic operating systemupstream description
-
Bifrost TypeSafe gateway route
A Go gateway provider that passes the native API through one-to-one, so the official SDKs work by changing only the base URL.
-
laya-mlx
Runs Laya's open-weight typed-decision checkpoints, not Jev, natively on Apple Silicon with MLX and no PyTorch or cloud API. Choice, score and noul answers come from one forward pass: 13.4 ms median for a short English decision, 7.4 ms multilingual, by the author's benchmark.written by a model
-
ax
DSPy-style framework for typed LLM programs, TypeScript first and also compiled to Python, Java, C++, Go and Rust. Its native TypeSafe client sends boolean, class and ordered-score signatures to Jev as noul, choice and score questions, and can keep each answer's probabilities and confidence.written by a model
-
memsearch
Shared semantic memory for AI coding agents such as Claude Code and Codex, kept as Markdown with a Milvus index. Optional Jev reranking sends the query and every candidate in one request, a noul per passage on whether it states what the query asks, and sorts results by those probabilities.written by a model
-
vellum-assistant
Personal AI assistant with eight kinds of memory that works across apps and chat channels. Its experimental TypeSafe provider sends the conversation to Jev as state and by default asks one noul, whether the latest message asks for a specific action, replying with the answers as JSON.written by a model
-
celesto
Secure and persistent computer for AI agents -- build your own Grokbot, and Muse.
-
laya
Node.js and TypeScript runner, via ONNX Runtime, for Laya, an open-weight model that copies Jev's System One interface: one forward pass answers choice, score and noul questions about a state, matching the Python reference to four decimals, about 140 ms for three questions on a warm Apple CPU.written by a model
-
kody
Cloud home for an AI agent's memory, keys, code and automations, served over MCP from Cloudflare Workers. On paid plans, when hybrid search returns an unclear pool, Jev scores each candidate on a four-level relevance rubric and weak ones are dropped; failure or low confidence keeps the hybrid order.written by a model
-
jev-chat-windows
微信(Windows 4.x)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
-
distill
Coding-agent harness and TUI for Grok and Codex subscriptions and any OpenAI-compatible model, built to spend fewer tokens. When a model tier's effort is left on auto, the default, Jev picks the reasoning effort for every call; the client reaches Jev through OpenRouter by default.written by a model
-
aiavatarkit
Toolkit for voice AI avatars in apps, phone lines, signage and the metaverse. Its Jev turn-end gate asks a noul on the transcript, whether the assistant should stay silent and leave the user's turn open, and maps that probability to a wait of 0.4 to 2 seconds; on any error the turn ends.written by a model
-
jev-skill
An agent skill plus CLI that validates all three primitives, requires explicit consent before a billed call, and forbids inventing output when simulating.
-
req_llm
Elixir library that gives LLM providers one request and response shape, built on Req and Finch. Its evaluate/4 sends named choice, boolean and score questions about a text or JSON state to Jev, directly or through OpenRouter, and returns each answer's probabilities; TypeSafe serves no chat.written by a model
-
jev-docs-zh
Jev 模型(TypeSafe AI)官方使用文档的中文翻译 | Unofficial Chinese translation of the official Jev (TypeSafe AI) docs — https://docs.typesafe.ai
-
jev-chat-jarvis-mac
微信消息意图识别悬浮窗(macOS):看屏 + 本地小模型判断意图和风险,再按话术生成回复候选。纯只读、不注入微信。
-
smithers
Smithers is an agentic workflow framework for defining workflows in simple TypeScript configuration files and executing them quickly, durably, and reliablyupstream description
-
jev-experiments
Nader Dabit's collection of small Jev experiments, one per folder: a commit reviewer, a shell guard, a send guard, a log sentinel, instant search, reranking, a voice-turn detector and more.
-
third-hand
macOS assistant you chat with: tag an app, such as @Spotify, and it reads the app's accessible controls, clicks and types. ChatGPT plans the steps; a Jev choice resolves each step's label to one on-screen element, or none, and one request can also ask whether the task is done and the next operation.written by a model
-
typesafe-mcp
The easiest first step once you have a key: registers Jev into Claude Code, Claude Desktop, Codex and Pi with one command.
-
crush-monitor
Local web tool that analyses a pasted WeChat chat with a crush. For each message Jev weighs 12 emotions and 35 intents, showing the top three of each, and it scores the affection signals and your own replies, graded SSS to D. Text LLMs can stand in for Jev.written by a model
-
pi-fabric
Runtime that gives the Pi coding agent one programmable tool for tools, MCP servers, agents and workflows, run as checked TypeScript or sandboxed Python. Its Jev skills add typed judgments: a foreman that watches coding work each turn and flags blockers, and bounded loops such as ticket triage.written by a model
-
orchestkit
Claude Code plugin of 110 skills, 36 agents and 163 hooks for review, testing, security and memory. Optionally, Jev is a second classifier of a session's work category: one choice over 9 categories from the git branch and first prompt; the cited script runs it on 150 held-out sessions.written by a model
-
laya-ultrafast
Port of jev-ultrafast, a browser agent, that makes its decisions on a Mac with Laya, an open-weight model, instead of Jev: about 33 ms per decision on an M1 Max. With DECISION_MODEL=typesafe it keeps the original Jev policy: in one request, a choice of operation and a choice of its target.written by a model
-
awesome-jev (fatwang2)
A sibling directory whose submissions are reviewed by Jev itself, with a notably thorough list of multi-language community clients.
-
openwhisper
Local speech-to-text, dictation, and meetings with Whisper and OpenAI API. Optional Windows x64 engines: Parakeet, Qwen3-ASR, Nemotron Streaming, and Moonshine.
-
djev-spark
Container recipe that serves DiffusionGemma 26B-A4B, not Jev, on a DGX Spark through vLLM, with the djev server in front answering Jev-style noul, choice and score requests on /v1/systemone. The cited probe times one choice and one noul over documents of growing length.written by a model
-
Jevmind
An agent's decisions pulled out of its prose into one place: typed answers with a confidence, behind a gate written in code, sealed in a ledger, graded and learned from. Runs offline on a local rules brain; Jev can be swapped in with one flag.
-
runline
Code mode for agents: 204 bundled plugins turn APIs and host functions into actions that agent-written JavaScript calls in a QuickJS sandbox. Its TypeSafe plugin sends a state and choice, score or noul questions to Jev, and first checks the request against Jev's token ceilings.written by a model
-
webctl
Web search CLI for agents: it queries 3 search backends, has Jev score each result against the query and goal (on topic, from a good source), keeps the best and dedupes them with MinHash and Jev. With --filter-chunks it scrapes the top pages and keeps only the chunks Jev judges relevant.written by a model
-
Building with TypeSafe Jev
Unofficial agent skill that teaches coding agents to design for Jev rather than treat it as a chat model: typed questions, thresholds on calibrated confidence, and links to 150+ community projects sorted by shape with a code sketch for each. Installs in Claude Code, Codex and others.written by a model
-
dasheng
Read-aloud trainer for English: a streaming ASR model transcribes, alignment picks the words that differ, and Jev judges only those, a noul on whether it is the same word, a choice of error type, and a 1–5 score of whether the sentence kept its meaning. Pronunciation is not scored.written by a model
-
Jev
Unofficial Jev showcase and lab: a smart-home demo that asks 13 questions in one call and lets code flip the lights, a playground for choice, score and noul with a policy in code, a cost lab, and a JS package that routes models and screens inputs for agent loops.written by a model
-
jevmem
Project memory for Claude Code, Cursor and Codex: a Jev noul asks if a decision is worth keeping, kept ones go to JEVMEM.md and return when relevant, and edits and commands are checked against them. Author's runs: Claude followed them in 66 of 72 sessions, 28 without, 67 with a hand-kept CLAUDE.md.written by a model
-
jev-voice
Voice control for the Mac: whisper.cpp transcribes locally, then one ~250 ms Jev request turns the command into an action and arguments picked from candidates code built, such as installed apps. Multi-step tasks run an agent loop on the Accessibility tree, capped at 40 actions and 80 Jev calls.written by a model
-
jev-foundation-models
Swift 6 bridge letting Apple's Foundation Models call Jev and similar models: a @Generable type's Bool fields become nouls, enums choices and @Guide ranges scores, answered against app state with probabilities. Jev's API is the default; on-device Core ML and self-hosted backends are optional.written by a model
-
agent-router
CLI that picks the coding agent, model and effort for a task and starts it in Herdr: fixed rules filter by quota and a 40% reserve on shared accounts, Jev ranks the rest and picks the effort, and an unsure pick on a consequential task goes back to you. Cursor, Claude Code, Codex or OpenCode.written by a model
-
stanley-code
Code-change checker CLI: a Jev choice routes each request to a workflow or cannot_tell; each workflow is code asking Jev small questions about bounded evidence, such as whether a hunk belongs to the task or a test got weaker. Other requests go to a coding agent, and a workflow for them is drafted.written by a model
-
laya-vs-jev
Side-by-side T-Rex runner where Laya, a local model on Apple Silicon via MLX, and Jev over TypeSafe's API each choose the next move from those a physics planner labels; safety checks can override a stale or unsafe pick unless --unassisted is set. Latency, streaks and crash replays are shown live.written by a model
-
jevchat
Turns Jev into a chatbotupstream description
-
jev-leftpad
Left-pad strings with TypeSafe AI's Jev. For reasons.upstream description
-
captaincore
👨🏽💻 CaptainCore is a command line application for automating WordPress maintenance.upstream description
-
jev-playground
A playground for calling Jev from MoonBit, with a CLI per question type (score, choice, noul) and measured reports on which composition patterns win.
-
openthai-systemone
OpenThai-SystemOne: open Thai + English System One decision model (0.8B, 256-way slot head, Apache-2.0)upstream description
-
dspy-typesafeify
Add a decorator for dspy Signatures that automatically uses TypeSafe where relevantupstream description
-
jev-rules
Jev picks which of your rules apply to each prompt, so Claude only sees the ones that matter.upstream description
-
jev-mcp
MCP server that puts TypeSafe Jev on the coding loop in Cursor, Codex, and any MCP clientupstream description
-
jeview
An unofficial local visualizer for Jev (TypeSafe): a live view of every call your code makes. Not affiliated with TypeSafe AI.upstream description
-
jev-paint
Use Jev to make art!upstream description
-
james_library
R.A.I.N. Lab is an experimental scientific-agent architecture that separates fast local judgment, independent probabilistic evaluation, multi-agent deliberation, evidence, and authorization into distinct computational layers.🐙(Predates Karpathy's AutoResearch)earlier upstream description
-
jev-trades
Trading bot with the all new TypeSafe AI's first system one model named as Jevupstream description
-
call-coach-ai
Jev powered call coachupstream description
-
jev-judge-mcp
Typed judgment tools for MCP agents. TypeSafe's Jev model as verify, screen, find, classify, rerank, decide, compare, extract, review, gate, and score: the model judges, policy decides auto, review, or escalate.upstream description
-
jevocks
Everyday Stocks Status with Jevupstream description
-
jevgraph
Evidence-backed knowledge graph construction with typed Jev relation decisionsupstream description
-
jevify
An agent skill to discover TypeSafe Jev opportunities, design typed questions, and learn from recent community experiments.upstream description
-
jev-tree
Jev-native probability tree and graph runtime for verifiable multi-step decision making.upstream description
-
cultivar
Use cultivar to test your Agent Skills and Docs by running them in sandboxes, and across different agents.upstream description
-
jev
A claude code plugin for jevupstream description
-
ask-jev-skill
Skill for Hermes, and other agents, to ask typesafe's jevupstream description
-
jev-cli
Command-line tool for TypeSafe AI's Jev model. Ask yes/no, multiple-choice and rubric questions about any text and get calibrated probabilities back. Answers become exit codes for shells and CI, JSON for scripts, and MCP tools for AI agents.upstream description
-
jev-spring-boot-starter
A simple Spring Boot 4 starter for TypeSafe Jev using Spring MVC and RestClientupstream description
-
clash-jev
A Clash Royale bot with no trained policy: Jev (TypeSafe System One) makes every decision from the live game stateupstream description
-
jev-register-tool
TypeSafe(Jev / System One)申请 → 确认邮件 → 获批 → 注册 → 建 API Key 全链路工具,纯 HTTP 无浏览器upstream description
-
st-jeved
SillyTavern extension that measures each reply and instructs the narrator only when a rule matches.upstream description
-
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 / PlayJevupstream description
-
is-jeven
Is it even? Ask Jev.upstream description
-
jev-skill-suggester
用 TypeSafe Jev 推荐已安装 Skill / Bounded installed-skill recommendations with TypeSafe Jev. Python CLI, Codex skill, bilingual docs and live examples.upstream description
-
duckdb-jev
DuckDB extension: typed Jev answers as real SQL typesupstream description
-
jpp
J++: an experimental language with standalone source and a Rust runtime. Compose questions and methods. 独立源码,组合问题与方法。upstream description
-
laya-vs-jev-arena
Laya (open source, local) vs TypeSafe Jev (API): two AI models race in Snake and fight in a Mortal-Kombat-style arena. Every move is a real model decision.upstream description
-
llm-typesafe
LLM plugin for accessing Jev and other TypeSafe AI modelsupstream description
-
jevvy
Jev-powered plugins for coding agentsupstream description
-
loki
The agent that evolves with you 𖤍upstream description
-
jev-trip
Two Minds, One Trip. https://jev-trip.vercel.app/earlier upstream description
-
typesafe-playground
Community TypeSafe AI playground: 110 use cases, games, dilemmas and model challenges, with editable prompts, A/B comparisons and a mobile-friendly UI.upstream description
-
jev-recipes
80+ composable TypeScript recipes powered by Jev for AI agents, retrieval, answer verification, and conversation workflows.
-
jevloop
The agent loop where decisions don't cost a large language model call. Zero deps, runs offline, no API key needed.earlier upstream description
-
open-spark-jev
Open-source, local decision models inspired by TypeSafe’s Jev and System One - built on Qwen3 for NVIDIA DGX Spark.upstream description
-
jev-to-answer
A Book of Answers toy: ask a question and Jev picks the answer.
-
jev-canvas
Draw on a tldraw canvas with your voice and a pointing finger. Jev (TypeSafe System One) decides action, target and place in ~350 ms per spoken word.upstream description
-
jev-tetris
Jev play Tetris in real-time against other AI modelsupstream description
-
Jev
Six live, side-by-side demos of Jev deciding while an LLM writes the words and code owns the control flow; the LLM side switches between Claude and Kimi models.
-
jev-skill-router
Claude Code hook asking Jev which installed skill fits each prompt: a choice ranks all skills beside three noul gates, then a choice and a noul per skill judge the top three, none suggested under 0.30. Shadow mode only logs; in the author's week, 28 of 539 suggestions were followed by that skill.written by a model
-
jot
The first general-purpose System One agent for Jevupstream description
-
jev-vs-ml
Jev-vs-MLupstream description
-
switchboard
An open-source, model-agnostic decision router for Claude Code and Codex.earlier upstream description
-
jev-chat-windows-deepseek-jev
Windows 微信回复助手:DeepSeek 官方生成话术,TypeSafe JEV 官方判断排序,支持可取消的 3 秒自动发送。
-
jev-studio
if you're experimenting with jev it will be easier from hereupstream description
-
jev_stock
An experimental JEV-powered framework for forecasting short-term stock price direction from structured market data.upstream description
-
jevernetes
Live Kubernetes log analysis, contextual investigation, and agent handoff powered by Jev.upstream description
-
jevthoven
AI Music (MIDI) generator powered by Jevupstream description
-
laya-jev-lab
Independent measurements of typed-decision models: Jev (TypeSafe API) vs Laya (open weights), and a local-first cascade that matches Jev's accuracy at 1.8x the speedupstream description
-
typesafe-playground
Interactive experiments with TypeSafe Jev, from support routing to 3D driving simulations with real AI decisions and visible sensor inputs.upstream description
-
awesome-jev
TypeSafe System One / Jev community directory — GitHub projects & posts around typed decisions (typesafe.ai)
-
Jev_Ontology
Trying to combine Jev with ontologyupstream description
-
jevtown
Jevtown: a social network where people write and 10,000 AI personas reactupstream description
-
typesafe-skill-router
TypeSafe (Jev) skill routing for Hermes Agent: names the one skill worth loading, before the model call. Opt-in, stdlib only, ~$0.001 per routed turn.upstream description
-
jev-chat-for-twitch
Filter any live Twitch chat with Jev: a bring-your-own-key Chrome extensionupstream description
-
jev-blindspot
A side-panel assistant that finds the blind spots in your prompts. For Claude Code and Codex CLI.earlier upstream description
-
pi-quiet-ask
TypeSafe Jev as the pi coding agent's quiet decision layerupstream description
-
trade-jev
Backtest Jev (TypeSafe) as a BUY/SELL/HOLD trader on NQ L10 order-book dataupstream description
-
jev
Typed decisions from the shell: an unofficial stdlib-Python CLI and Agent Skill for TypeSafe's Jev model, via the TypeSafe API (default) or OpenRouter. Yes/no, choice and ordinal scores with calibrated probabilities, semantic grep and batch mode.upstream description
-
jev-realtime-trading
Paper trading agents on a live tape, decided every second by TypeSafe's Jev (System One). Electron desktop app.earlier upstream description
-
jevplayspokemon
Jev plays Generation 3 Pokémon via Showdown and a real FireRed ROM.upstream description
-
jeq
What happens when jev meets jq? Intelligence you can pipe for quick experimentation and scriptsupstream description
-
jev_project_context
Evidence-first long-term experiment memory skill for AI coding agents, with optional Jev decision-model layersupstream description
-
typesafe-chess
Chess where both players are TypeSafe's Jev model: every move is a typed Choice decisionupstream description
-
jev-search
Experimental semantic line search with TypeSafe Jev via OpenRouter. Python CLI with no runtime dependencies.upstream description
-
soupbase
Jev x 海龟汤upstream description
-
jcm-router
Local proxy that picks the Claude model and effort per message using TypeSafe Jev. Routes subagents, leaves your cached main chat alone.upstream description
-
jev-minesweeper
TypeSafe Jev (System One) にブラウザ上のマインスイーパーを解かせるデモupstream description
-
secondlayer
Decoded Stacks data in your own database. Self-hosted.earlier upstream description
-
ai-elo-ranker
High-speed recursive AI Elo tournament engine powered by Jev and Swiss matchmakingupstream description
-
cairn-jev-lab
Test what your AI should remember. An experimental, source-aware memory admission evaluator powered by Jev, with editable cases and inspectable results.upstream description
-
jev-grand-prix
An F1 racing game where TypeSafe's Jev picks the racing line and the pedals, and learns each corner's limit between lapsupstream description
-
jev-mcp
MCP server exposing TypeSafe Jev as typed, calibrated judgment tools: classify, score, check, batched ask. Ships as a Claude Code plugin.upstream description
-
jev-t-rex-runner
Chrome dino game played by Typesafe AI Jev modelupstream description
-
jev-yt-time-saver
A Chrome extension that covers distracting YouTube videos with Jev. Show anyway whenever you want.upstream description
-
jev_jsonschema
Run a JSON Schema through TypeSafe's Jev API, and get JSON back.upstream description
-
jevtest
情緒測謊器:嘴上說「好」,心裡真的好嗎?用 TypeSafe Jev(System One 模型)透過 OpenRouter 即時判斷,並與一般 LLM 對照upstream description
-
mcts-agent
Discriminative Monte Carlo Tree Search using TypeSafe Jev System One Primitives and Gemini
-
rubikjev
Challenge the Jev's intelligence in Rubik Cube puzzlesupstream description
-
typesafe-ui
shadcn-style reusable components and blocks for using TypeSafe AI.upstream description
-
jev-bot
Self-hosted Jev decision workbench and Feishu bot: automatic choices, probabilities, and experimental word/character writing.upstream description
-
jev-plays-pokemon
Playing Pokemon Red using TypeSafe Jevupstream description
-
jev-wingman
基于 Jev 的聊天决策辅助,不挑 App(QQ / 微信 / 飞书皆可)· An on-device chat co-pilot that returns typed verdicts instead of prose, built on Jev
-
n8n-nodes-jev
Helper n8n community node for Jev by TypeSafe. Classify, route, and score text with questions you define, and get a probability for every answer so unsure items can go to review.upstream description
-
your-signal
Open-source BYOK Chrome extension for personal, reversible X timeline filters.upstream description
-
btc-jev-signal
Experimental multi-horizon BTC signal generator using TypeSafe Jev probabilities and Binance market data.earlier upstream description
-
decisions-judge-mcp
Typed decisions for AI agents as an MCP tool: yes/no probability (noul), choice, and score in one fast request. Backed by the TypeSafe System One model.upstream description
-
jev-2048
An instrumented 2048 web lab where every move is a Jev (TypeSafe AI System One) Choice, with no heuristic fallback | 用 Jev 决策模型驱动每一步的 2048 网页实验台,概率、置信度、延迟与成本全部摊开可见,且刻意不做启发式兜底upstream description
-
jev-arena-nanojev
完全本地的 NanoJev 网格决策游戏实验场,支持中文 Pygame、多关卡与 GTX 1660S 训练upstream description
-
jev-system-one
A polished OpenAI + TypeSafe Jev terminal interface for answers with transparent decision reportsupstream description
-
jevopt
Making intelligent compiler optimisation decisions with Jevupstream description
-
jevtok
Exact token counting and request-cost prediction for TypeSafe's Jev (tiktoken-style)upstream description
-
alphaoptimizer
Jev-powered output optimization for Codex, built to keep large tool results concise and usable.upstream description
-
auto-mode-for-paseo
A Paseo provider that uses TypeSafe Jev to route each Codex turn.
-
jeff-cli
jeff, a Go CLI for Jev: from a shell script, give it state and a question with fixed answers and get calibrated probabilities back, with a rank command and meaningful exit codes.
-
jev-chat
A chatbot from typed Jev decisions: hierarchical speculative decoding over System One probabilities.upstream description
-
jev-connector
WordPress connector for TypeSafe Jev: typed, confidence-scored answers your code can branch on.earlier upstream description
-
jev-grug
Helping JEV speak <3upstream description
-
jev-x-kit
Offline $0 decision layer for coding agents: Choice/Score/Noul primitives, BELKI confidence gatekeeper, ultra-planning, red-teaming, research and RLVR self-improvement -- as an MCP server + CLI + Claude Code skill.upstream description
-
jevseek
DeepSeek proposes the next token, TypeSafe's Jev chooses it: a decision model used as a samplerupstream description
-
jevslop
Jevによるnote記事のAI Slop判定サイトupstream description
-
new-api-plugin-typesafe
TypeSafe AI System One (Jev) task plugin for QuantumNous/new-api — native /v1/systemone, synchronous evaluation, token billingupstream description
-
pydantic-jev-examples
Pydantic AI capabilities made stronger with Jev: small runnable demos, one file eachupstream description
-
rust-sysone
System One TypeSafe AI Rust Client (unofficial)upstream description
-
skill-router
Tell Claude Code which installed skill a session needs, using Jev (TypeSafe AI) for the decision and skills.sh for discovery.upstream description
-
sqlite3-jev
SQLite extension that calls TypeSafe Jev (or tensai serve) from SQLupstream description
-
system-one-chess
Chess against Jev, TypeSafe AI's System One model, through OpenRouter. Built with the pico framework.upstream description
-
typesafe-ai-playground
A playground for experiments around Jev, TypeSafe's System One model.upstream description
-
awesome-jev-use-cases
Explore real-world use cases and projects built with TypeSafe AI's Jev: content moderation, AI agents, model routing, and semantic search. Curated by SeeAPI.upstream description
-
dbt_jev
use jev in dbtupstream description
-
emoji-jev
Emoji autocomplete at the speed of typing. TypeSafe AI Jev on a Whop-hosted TanStack Start app.upstream description
-
everything-about-jev
tell you everything about jev,TypeSafe AI's System One model for typed decisions.upstream description
-
hermes-jev-curator
Typed Jev relation governance and safe archive plans for Hermes Curatorupstream description
-
jev-cli
CLI for TypeSafe AI's Jev evaluation model — typed questions in, structured JSON answers outupstream description
-
jev-codex-router-skill
Portable Codex Skill for Jev model and reasoning-effort routing, with safe installation and Chinese usage guidesupstream description
-
jev-freeform
An observable raw-character chat experiment powered entirely by TypeSafe Jev Choiceupstream description
-
jev-gomoku
弈瞬:双 Jev 五子棋九宫格输入实验台,逐手查看模型决策,支持真实对局回放与实时对战。upstream description
-
jev-mcp
An eval-first MCP server for TypeSafe's Jev, a System One model that returns typed judgments (noul, choice, score) with probabilities instead of generated text.upstream description
-
jev-mcp
MCP server for typed decisions with Jev / System One via OpenRouter or TypeSafeupstream description
-
jev-pii-checker
CLI that finds PII in text with TypeSafe Jev: presence, sensitivity, and located spansupstream description
-
jev-resume-disqualifier
Jev Resume Disqualifier: Sub-25ms automated resume knockout engine powered by TypeSafe Jev System One decision intelligence. Eliminates 80% of unqualified applicants with deterministic date math & EEOC-safe rejection notices.upstream description
-
jev-skills
Practical agent skills and examples for building with Jev. API setup, routing, ranking, and evidence checks.upstream description
-
n8n-nodes-typesafe-jev
n8n community node for TypeSafe Jev structured AI decisionsupstream description
-
research_desk
TypeSafe Jev demonstration for new analyzation — experimenting with Jev for fast analysis of news and tickersupstream description
-
should-ai-kill-us-all
We ask Jev, TypeSafe AI's System One model, whether AI should kill us all. Every ten minutes. Using the actual headlines.upstream description
-
systemone-lite
Toy local System One–style decision API (Jev-shaped). Not affiliated with TypeSafe.upstream description
-
typesafe-assist
Home Assistant Assist conversation agent powered by TypeSafe's Jev (System One) modelupstream description
-
typesafe-jev-mcp
MCP server exposing TypeSafe's Jev model as a typed evaluate tool.upstream description
-
agent-jev-tetris
using the new model JEV to play the game tetrisupstream description
-
ailerix
Type-safe model router. Jev (System One) banks each request to a typed catalog route.upstream description
-
barrunto
A Chrome extension that brings TypeSafe's Jev to X.com to analyze posts as you browseupstream description
-
bes-kelime-jev
Ne yazarsanız yazın, beş kelimeden biriyle cevap veren sohbet botu. Kelimeyi TypeSafe AI'ın Jev evaluation modeli seçer.upstream description
-
codex-jev-preflight
Fail-open Codex UserPromptSubmit hook that injects TypeSafe Jev pre-task routing metadata.upstream description
-
git-jev-stage
Select Git changes for staging with a plain-language description.upstream description
-
got-jev
Jev (TypeSafe AI) PoC through Game of Thronesupstream description
-
ha-conversation-jev
Home Assistant custom component: Conversation agent with Jev fast-path + Grok fallbackupstream description
-
jev-broadcast-lab
Testing Lab for Jev AIupstream description
-
jev-ci-selector
CI task selection for GitHub Actions with Jev and a pure policy engine; current selection is applied by default, with an explicit shadow mode for observation.
-
jev-cloud-quiz
A demo in which Jev judges, with probabilities, which of the three big clouds a feature name belongs to.
-
jev-cvss
Fast CVSS scoring from vulnerability descriptions using Typesafe Jevupstream description
-
jev-demo
Jev (TypeSafe System One) 学习与实测:概念文档 + 5 个可运行 demo + 可复现压测。实测 jev-1.13.0:扇出几乎免费,40 问与 1 问等延迟。upstream description
-
jev-games
Visual Jev lab for multiple games and emulator platformsupstream description
-
jev-mcp
MCP server wrapping TypeSafe's Jev System One models — typed noul/choice/score judgments for AI agentsupstream description
-
jev-mcp-spring
Java/Spring Boot MCP server for TypeSafe Jevupstream description
-
jev-playground
Can a System One model steer music? Jev picks the plan (enums only); code renders sheet, audio and MIDI.upstream description
-
JevTape
A record-and-replay tool for Jev decisions: a CLI, a local proxy and JSON tapes, with replay fully offline.
-
magic-jev-ball
A Magic 8 Ball that asks Jev instead of picking at random. Convex + AI Gateway + three.js.upstream description
-
pi-agent-foreman
Send Pi agents back to work when they stop before the job is done.upstream description
-
risc-jev
I tortured Jev into being a RISC-V CPU.upstream description
-
tempo-jev-demo
A natural-language task workspace comparing performance across AI models (TypeSafe's Jev, GPT-5.6 Luna, and Gemini 3.8 Flash)upstream description
-
typesafeai-cli
Give your AI agent a CLI companion who has access to TypeSafe AI's Jev.upstream description
-
wellposed
Lint your jev requests before they come back confidently wrong.
-
askjev
Unofficial MCP server for Jev (Typesafe AI)upstream description
-
harden-jev-decides
JEV picks which stream idea becomes the live MVP. TypeSafe System One decision board.upstream description
-
jev-acp
Use Jev typed decisions from any ACP (Agent Client Protocol) client or IDEupstream description
-
jev-demo
A zero-dependency web bench for TypeSafe Jev with three presets — browser actions, intent recognition and agent context pruning: send state and typed questions, get calibrated structured answers.
-
jev-eyes
Give Jev eyes — honest, local image perception for TypeSafe's text-only System One model. OCR + spatial layout → Jev state. CLI, MCP server, agent skill.upstream description
-
jev-measured
Measured cost, latency and raw output from the live Jev API (TypeSafe AI System One model) across 8 use cases — reproducibleearlier upstream description
-
jev-pick-and-place-study
A small reproducible MuJoCo pilot comparing Jev, Claude Haiku, and reactive rules for pick-and-place.upstream description
-
jev-playground
A small Next.js app for experimenting with TypeSafe AI's Jev model (System One)upstream description
-
jev-practice-speed
A WebGL demo where you play the card game Speed against a CPU whose brain is TypeSafe AI's Jev. The whole point of the app is to measure and show Jev's decision speed and decision accuracy in real time.upstream description
-
jev-snake
An experimental Snake environment where the game engine owns deterministic rules and TypeSafe AI's Jev makes the movement decision from structured state on every tick.upstream description
-
jev2048
Let Jev (TypeSafeAI) solve 2048upstream description
-
jevals
Local evaluation workbench for TypeSafe Jevupstream description
-
jevcode
JevCode: technical solutions and best practices for Jev (TypeSafe System One), with a companion website.
-
Jevometry
an Information-Geometric Analysis Toolkit for any System-one (Jev, Jevlike) agent systemsupstream description
-
openpoke-meets-jev
Open source implementation of Pokeupstream description
-
r2r-jev
Persistent governance for AI agents — turn Jev judgments into replayable relation state with R2R.upstream description
-
second-thought
Learning infrastructure for typed probabilistic decisions from System One models (Laya, and typed-decision providers you bring yourself).upstream description
-
typesafe-jev-examples
Worked examples for TypeSafe's Jev System One decision model, runnable today through OpenRouterupstream description
-
typesafe-jev-tools
A Claude Code hook that asks whether the decision you are writing needs a model at all. Includes a measured 149-row comparison of TypeSafe Jev against Claude Haiku 4.5.upstream description
-
aegis: TypeSafe as a first-class provider
A personal Rust AI gateway with a TypeSafe provider, usage extraction and alias resolution tested against real response bodies.
-
AskJev-MCP
MCP server for TypeSafe's System One API (Jev): typed choice/noul/score judgments with calibrated probabilities and confidenceupstream description
-
beatjev
try to beat jevupstream description
-
cartshield
CartShield — SMB checkout fraud disposition via TypeSafe Jevupstream description
-
commentcop
Put your code comments on trial. Powered by Jev.upstream description
-
cyber-breach-jev
Cyber-Breach: The Jev Protocol - A tactical cyberpunk arena combat game powered by TypeSafe AI Jev System One decision modelupstream description
-
dgp
Decision Graph Protocol (DGP) by Numerous ApS: open-source contracts for decision-based AI agents, TypeSafe Jev orchestration, guarded actions, and assessment batching.upstream description
-
extremely-specific-council
Twelve members. Zero qualifications. A playful TypeSafe AI council with animated votes, inspectable decisions, and shareable verdicts.upstream description
-
financialpredictionjev
Using Jev to test how well it predicts financial markets(just like most llms as of september 2026, it doesnt do that good)upstream description
-
frost
A flexible and configurable CLI model router using TypeSafe Jev.upstream description
-
functions
👷 Cloudflare Workers for the TrainLCD mobile app.upstream description
-
hiresignal
HireSignal — resume first-pass fit+interview via TypeSafe Jevupstream description
-
jev-anotacao-sentencas
Jev (TypeSafe) vs. Gemini 3.8 Flash vs. GPT-5.6 Luna na anotação estruturada de sentenças do TJSP: qualidade, tempo e custoupstream description
-
jev-calculator
A probabilistic AI calculator powered by Jev.
-
jev-evaluation
An adversarial evaluation of TypeSafe's jev decision model: nine experiments and 28 predictions fixed before any data was collected. 123,805 requests, $12.69.upstream description
-
jev-torneo-animales
Winner-stays-on animal tournament refereed by Jev (TypeSafe System One): a local game to feel how fast typed decisions are. UI in Spanish.upstream description
-
jevscope
Local-first visual decision debugger and regression testbench for TypeSafe AI Jevupstream description
-
labs
Small, independent projects for experiments, research, and investigations.upstream description
-
mcpmatch
Match user goals to MCP catalog (two-stage) via TypeSafe Jevupstream description
-
mimicry
Rewrite AI drafts in your own voice with a bounded TypeSafe feedback loop.upstream description
-
n8n-nodes-typesafe
Typesafe AI Node for N8Nupstream description
-
open-jev-bridge
MCP plugin to connect jev-like system one API (local hosted or typesafe jev) to codex and claude code for decision tasks like compaction, verification judgement, etc.upstream description
-
pi-typesafe
Pi coding-agent extension built on the TypeSafe AI System One API (Jev)upstream description
-
pong-jev
TypeSafe's Jev plays Atari Pong. One typed Choice question per frame, no coordinates sent to the model.upstream description
-
river-run-typesafe
River shooter game in Python, inspired by Atari's River Raid, played by a TypeSafe AI pilotupstream description
-
scam-shield
Scam text message filter powered by TypeSafe's Jev modelupstream description
-
search-function-test
A test project based on Jev AI, the goal is to build a search function for a blog/article website that has 100s of articles to search from, So the user can actually use the search as chat to question anything and find related answers/articlesupstream description
-
typesafe-comment
Small Python package that uses typesafe.ai to evaluate code comments on certain heuristicsupstream description
-
typesafe_chess_eval
An evaluation of typesafe AI chess. As it turns out, the AI isn't doing really well even though chess is not a particularly open-ended game. Still, it's only a prototype and this probably wasn't optimzied for games.upstream description