Projects built with Jev

browser-use/jev-ultrafast

Fast browser agent from Browser Use. Jev decides each step and which element to act on; a small model is called only when text needs to be typed. The authors report a full Google Flights search in about 7.1 seconds.

tamaratran/fast-jev-compaction

Claude Code plugin that replaces the compaction summary with Jev decisions. Every tool call and result is scored; stale ones are dropped or truncated, and everything kept stays verbatim.

lahfir/agent-desktop

Desktop automation over OS accessibility trees. The feat/jev-desktop-loop branch uses Jev to choose which control to operate and which action to take.

TheoLeeCJ/openjev

Jev-style decisions from a frozen 4B model on a single RTX 3090, with a browser demo.

vinnylarouge/jevlike

Train a small model that scores a changing list of text options in one pass.

jarrodwatts/jev-trader

Asks Jev buy or sell on every Monad block and places real post-only limit orders on the Kuru MON-USDC book. Not financial advice.

fhshaik/typesafe-mario

Jev picks NES controller inputs for Super Mario Bros. from emulator state, with no screenshots.

thruwire/foreman

Puts Jev as a fast supervisor above slower coding agents such as Codex, starting from a ticket, spec, or bug report.

devagrawal09/jev-review

Staged code-review workflow with a local dashboard. Jev triages correctness, security, reliability, compatibility, and test risk, then hands the important findings to a heavier model.

awlevin/typesafe-computer-use

macOS computer use without sending screenshots to a large model. The screen is read deterministically and Jev picks the next action.

TianyuCodings/NanoJev

0.6B parallel decision model with an end-to-end training pipeline.

gargpratyush/jev-router

Per-turn model routing for Claude Code and Codex. Simple work goes to the fast tier and difficult work to the strong tier.

NiazMorshed2007/jev-review

Local-first MCP server for continuous software-quality review by coding agents.

ekzhang/openjev-sglang

Jev-compatible API server running an open model on SGLang.

realZachi/pg-jev

PostgreSQL extension to filter, rank, and classify rows with plain-language conditions.

droidrun/mobile-jev

Phone automation from DroidRun. The demo sets up an Uber ride in about 21 seconds.

kitze/unclutter

Browser extension that removes page clutter using Jev and reusable template rules.

pithings/advocaat

Small type-safe TypeScript client for asking Jev about your data, with an agent skill for designing questions.

jkudish/jev-mcp

Proof-of-concept MCP server with ready-made tools for fact checking, prompt-injection detection, and semantic ranking.

RomanSlack/jev-drone

Camera-only quadrotor in MuJoCo with Jev making judgment calls at about 2.5 Hz.

DevMortimer/pi-warden

Guardrails for the Pi coding agent. Jev judges every write and edit against the rules in pi-warden.md.

dbreunig/building-with-jev-skill

Agent skill for writing programs that call Jev: question design, state structure, confidence thresholds, and diagnosing wrong answers.

itsmostafa/typesafe-mcp

MCP server that lets agents such as Claude Code, Claude Desktop, and Codex call Jev directly for Choice, Score, and Noul decisions.

ChetasLua/jevmeter

Scores every sentence in a video and renders a live Jev meter as a 16:9 edit.

standardagents/jevpilot

Three.js driving simulator with a Jev-powered autopilot.

jkudish/jev-browser

Headless browser automation through an MCP server, CLI, or library. Jev picks one action per step.

Dicklesworthstone/skillranker

Rust CLI that ranks which agent skills fit the next step from live session context, with Claude Code hooks.

realZachi/typesafe-adblock

Chrome extension that asks Jev whether each DOM element is an ad and removes the ones that are.

irfndi/prism-liquidity-agent

Liquidity-provision agent for Meteora DLMM. Jev judges toxic flow, market stress, and mean-reversion likelihood in shadow/advisory mode only, without driving trades. Not financial advice.

iammrduncan/typesafe-ai-benchmark

Compares LLM structured output with Jev on latency, cost, and judgment quality.

brainstormity/Jev-Moderation-Bot

Discord bot that filters spam and scam links in real time and escalates repeat offenses.

0xNatoshi/jev-codex-router

Picks the model, reasoning depth, and speed mode for every Codex turn based on how hard Jev judges it to be. The author reports about 60% lower cost when replaying 237 real turns.

devanshbatham/commit-miner

Classifies Git commits into bug fixes, security fixes with CWEs, and change types.

bnsd55/jevmlx

Jev-style typed decisions from local MLX models on Apple Silicon.

ellipsis-dev/blink

Semantic codebase search. At each directory level Jev ranks which files and folders are most likely relevant and sends more walkers there.

GhalebDweikat/winnow

Context sieve for Claude Code. Jev judges each tool result (Read, Bash, Grep output) for relevance before it enters the context window.

jexp/neo4jev

Navigates a Neo4j graph by having Jev score neighbouring relationships, then beam-searching for the most probable path.

phyous/tsai-sc

Jev plays the first StarCraft shareware mission, with its recorded action probabilities.

kieranklaassen/ruby_llm-typesafe

TypeSafe provider for RubyLLM 2.

AboveColin/HA-Jev

Home Assistant integration that turns Jev's answers about your house into entities.

AbdelStark/jev-benchmarks

Probability-aware evaluation: calibration, and how much work can be automated at a fixed error budget.

saibimajdi/typesafeai-dotnet-sdk

.NET SDK.

jomatsu/zod-jev

Adds Jev semantic checks to Zod 4 schemas. Zod validates the shape; Jev validates the meaning.

huntedman/JevLint

Checks code against conventions written in plain English, in a write-check-fix loop with your coding agent.

sharziki/semdecide

Unix-style CLI for typed semantic decisions: classify, score, filter, and guard inside shell scripts, CI, and data pipelines.

mrnugget/jev-shell-history

Fish-style zsh autosuggestions, ranked by Jev from your recent history.

reachjalil/jevlogs

Scores OpenTelemetry logs for diagnostic value, priority, and routing before expensive LLM analysis.

emrickgarrett/OneVOneJev

Three.js quickscope arena where Jev decides movement, aiming, ADS, firing, and jumping at roughly 9 Hz.

yusukebe/hono-jev-router

Experimental Hono router that matches requests against plain-language descriptions instead of paths.

devagrawal09/jev-code

Bounded Jev workflows for coding agents.

andrelandgraf/safer-with-jev

Neon Function proxy for the Neon AI Gateway. Jev classifies each request and routes it to the right downstream model.

AntonioCoppe/jev-harness

TypeScript library that wraps Jev answers in policies, confidence gates, shadow mode, and reusable recipes.

iamvatsalpatel/tiershift

Sends each LLM request to the cheapest model tier that can handle it and escalates on evidence.

FirasSX914/Janus

Measures Jev's calibration and confidence-based routing on Banking77 and Web of Science.

nshkrdotcom/typesafe_sdk

Elixir SDK for the System One API.

kgonia/typesafe-sdk-java

Zero-dependency Java client (Java 21+).

reachjalil/jev-tree

Recursive Choice over a taxonomy, for picking among more options than a single Choice question allows.

BYK/jev-mcp

Eval-first MCP server that focuses on knowing whether Jev's answers can be trusted for your task.

noelzappy/tripwire

Runs seven checks on every LLM response in one Jev call, as AI SDK middleware or an OpenAI-compatible proxy.

markjaquith/typesafe-ai-playground

Rust CLI of Jev experiments, including PHI detection, code-comment review, live tone analysis, and occupation and industry classification.

abhixhek/jevcal

Picks the confidence threshold that meets your accuracy target on your own data, and fails CI when a model update breaks it.

Shubham510/typesafe-go

Go client.

kunobi-ninja/kunobi-jev

Rust client, published on crates.io.

zampierid4p/n8n-nodes-typesafe-ai

n8n community node for asking typed questions inside workflows.

pZacca/askjev

MCP server published on npm, with setup instructions for Claude Code, Claude Desktop, Cursor, and Codex.

rashedInt32/jev-gates

Six calibrated gates for Claude Code (rules, scope, intent, done, claims, and commit honesty) that escalate but never approve.

aoprisan/jev-ts-repl

Terminal REPL for shaping System One requests before writing code. Simulates answers when no API key is set.

m0rphtail/triagedy

Security alert triage as a Unix filter: JSONL alerts in, typed decisions out.

colliber/duckdb-jev

DuckDB extension that returns Jev's answers as real SQL types.

ably-labs/jev-pong

Pong where the ball moves one step per model decision, pitting Jev against chat LLMs.

denikuchero/jev-chess-lab

Recorded chess experiments with a candid result: Jev on its own still blunders pieces.