Read & learn
Written guides.
Understand Jev, one idea at a time. Walkthroughs, recipes, and writeups — our notes first, the original next.
Explore the Jev ecosystem.
Learn, build, and play with Jev — from deep technical guides to creative experiments.
Watch & learn
Video guides.
See an idea take shape. Tutorials, demos, and deep dives, organized by topic and credited to their creators.
Make & explore
Experiments.
See what builders made with Jev. Games, tools, repositories, articles, and more — traced to their sources.
The directory03 / 03 · Experiments
Find something worth exploring.
Every category, from games and tools to repositories and writeups, traced to a primary source.
Same ideas.
Different paths.
This is the full registry, not just the featured picks. Figures like speed, cost, and stars are maker-reported or captured snapshots, not JevMade measurements.
309 experiments · showing 1–60
DeerFlow includes opt-in Jev extensions for text classification, prompt-injection screening, and pruning old tool results.
SOURCE SCREENSHOTEXPAND ↗
from the source
async def _probability(self, client: httpx.AsyncClient, key: str, excerpt: str) -> f... body = { "model": self.options.model, "state": {"content": excerpt}, "questions": {"injection": _question()},
Keep what you find. Notes stay private.
Oh My ClaudeCode can compare Jev with its existing coding-agent heuristics in shadow mode, then optionally let Jev make those decisions while retaining the original behavior as a fallback.
SOURCE SCREENSHOTEXPAND ↗
from the source
export async function queryJev( state: unknown, questions: JevQuestions, options: JevClientOptions,): Promise<JevResponse> {
Keep what you find. Notes stay private.
Oh My Pi lets coding agents use Jev for structured decisions such as choosing an option or scoring a candidate.
SOURCE SCREENSHOTEXPAND ↗
from the source
async #attempt<T>(path: string, body: string, key: string, signal: AbortSignal | undefi... const url = `${this.baseUrl}${path}`; const headers: Record<string, string> = { ...this.#headers, Authorization: `Bearer ${key}`,
Keep what you find. Notes stay private.
jcode lets its fast browser agent ask Jev to choose the next move from the actions currently on screen.
SOURCE SCREENSHOTEXPAND ↗
from the source
async fn decide(&self, request: &DecisionRequest) -> Result<Decision> { let body = request_body(request)?; let questions = body["questions"] .as_object() .context("Browser decision questions are missing")?
Keep what you find. Notes stay private.
OpenChamber asks Jev to choose a model for Auto sessions and, when enabled, flag risky tool permissions for human review.
SOURCE SCREENSHOTEXPAND ↗
from the source
export const createJevClient = ({ fetchImpl = fetch, timeoutMs = JEV_TIMEOUT_MS } = {})... /** Resolves to the parsed answers; throws with `status` on an HTTP error and `code: '... ask: async (request, token) => { const abort = new AbortController(); const timer = setTimeout(() => abort.abort(), timeoutMs);
Keep what you find. Notes stay private.
Sure can ask Jev to sort bank transactions into a family's categories and benchmark those choices against labeled examples.
SOURCE SCREENSHOTEXPAND ↗
from the source
def decide!(state:, questions:, model: "") raise Error, "No questions provided" if questions.blank? questions = questions.transform_keys(&:to_s) questions.each { |key, question| validate_question!(key, question) }
Keep what you find. Notes stay private.
Qlty Slop One asks Jev to score code-quality traits in source excerpts, then rolls those scores into per-file reports.
SOURCE SCREENSHOTEXPAND ↗
from the source
/// The validated answers for one excerpt, from the cache when the exact /// request was answered before. pub fn request(&self, state: &RequestState) -> Result<Answers> { let body = state.body()?; let key = canonical::digest(&body)?;
Keep what you find. Notes stay private.
A Claude Code plugin that uses Jev to decide which context survives compaction.
SOURCE SCREENSHOTEXPAND ↗
from the source
claude plugin marketplace add tamaratran/fast-jev-compactionclaude plugin install fast-jev-compaction@fast-jev-compaction
Keep what you find. Notes stay private.
Mediary Scout filters irrelevant and adult-content media before those candidates reach more expensive workflow stages.
SOURCE SCREENSHOTEXPAND ↗
from the source
if (score === undefined) return true;const band = classifyJevScore(score);if (band === "drop") return false;
Keep what you find. Notes stay private.
Ask what code does and `jg` returns file locations, reading leads and source excerpts for a coding agent.
SOURCE SCREENSHOTEXPAND ↗
Keep what you find. Notes stay private.
Pi Typesafe Compact asks Jev which parts of a coding conversation deserve space in the compacted summary.
SOURCE SCREENSHOTEXPAND ↗
from the source
const selectedUnits = evaluation.decisions .filter((decision) => decision.summarize) .map((decision) => decision.unit);
Keep what you find. Notes stay private.
DSCode asks Jev whether a coding agent's risky action should proceed, stop or receive a fuller review.
SOURCE SCREENSHOTEXPAND ↗
from the source
verdict = await jev.approval({ action, context, sessionId: req.agent.session.id, signal:...if (verdict !== undefined) { if (verdict.decision !== 'defer') return applyVerdict({ decision: verdict.decision, re...}
Keep what you find. Notes stay private.
Coder One uses Jev to steer and check bounded autonomous coding sessions.
SOURCE SCREENSHOTEXPAND ↗
from the source
let mut judge = policy .judge(jev_client, workdir.clone(), &state.issue, recorder.clone()) .with_deep(policy.deep_jev());judge.survey(&mut state).await;
Keep what you find. Notes stay private.
See where a Claude Code session made progress and where it went in circles, with Jev reviewing a redacted trace.
SOURCE SCREENSHOTEXPAND ↗
from the source
key: "progressingEfficiently", label: "Progress", question: "Did the agent make steady, meaningful progress toward the user's task... higher_probability_is_better: true, answer: MetricAnswer::Noul,
Keep what you find. Notes stay private.
A Hermes toolkit for model routing, retrieval triage, memory, skill selection, context trimming and interface decisions.
SOURCE SCREENSHOTEXPAND ↗
Keep what you find. Notes stay private.
An MCP server that lets a coding agent ask Jev whether a claim is supported, a page is worth reading, or its own patch is safe.
SOURCE SCREENSHOTEXPAND ↗
choice
How does the evidence relate to claim `claim0` (`<claim text>`)? · supports / contradicts / says_nothing
Keep what you find. Notes stay private.
A software-factory foreman built around TypeSafe's Jev model.
SOURCE SCREENSHOTEXPAND ↗
from the source
TYPESAFE_API_KEY=your-key-here
Keep what you find. Notes stay private.
Tunes Jev decision functions from labelled examples by surfacing uncertain cases and proposing revised definitions through GEPA.
SOURCE SCREENSHOTEXPAND ↗
Keep what you find. Notes stay private.
A staged Jev code-review workflow with a local dashboard.
SOURCE SCREENSHOTEXPAND ↗
from the source
Noul risk matrix -> Choice + Score file profiles -> Choice evidence selection -> Choice mechanism classification -> Score severity
Keep what you find. Notes stay private.
skillbox is a self-hosted, versioned Agent Skills library with MCP access, scoped clients, and optional Jev recommendations.
SOURCE SCREENSHOTEXPAND ↗
Keep what you find. Notes stay private.
Audits each LLM code-review finding with one credibility question before deterministic CI rules decide whether to fail.
SOURCE SCREENSHOTEXPAND ↗
from the source
"questions": { QUESTION_ID: { "type": "noul", "question": VERIFY_QUESTION, "instructions": VERIFY_INSTRUCTIONS,
Keep what you find. Notes stay private.
Suggests a safe moment to compact an agent session by combining two Jev judgments with context usage.
SOURCE SCREENSHOTEXPAND ↗
Keep what you find. Notes stay private.
Autopilot's /jev command asks typed yes-or-no, choice or score questions about the current coding session.
SOURCE SCREENSHOTEXPAND ↗
from the source
const answer = await askJev(apiKey, question, { context });const presentation = presentJevAnswer(question, answer);setResult(eventKey, { kind: "jev", result: presentation.text,
Keep what you find. Notes stay private.
Adds typed tool, skill, owner, and quality decisions to agent harnesses without taking over their planning or execution.
SOURCE SCREENSHOTEXPAND ↗
Keep what you find. Notes stay private.
A Claude Code router that uses Jev to choose the cheapest suitable model for each task.
SOURCE SCREENSHOTEXPAND ↗
from the source
git clone https://github.com/gargpratyush/jev-router.gitcd jev-routernpm installnpm linkecho "JEV_API_KEY=..." > ~/.jev-router.env
Keep what you find. Notes stay private.
Filters oversized Bash output into verbatim chunks before Claude Code or Codex has to read it.
SOURCE SCREENSHOTEXPAND ↗
Keep what you find. Notes stay private.
A local-first MCP plugin for continuous software-quality reviews by coding agents.
SOURCE SCREENSHOTEXPAND ↗
Keep what you find. Notes stay private.
Medical MCP asks Jev to rank fetched clinical literature against a research question.
SOURCE SCREENSHOTEXPAND ↗
from the source
const applied: PolicyDecision = applyPolicy(judged.scores, row.evidence);const ranked = toRankedHit(row.hit, applied);
Keep what you find. Notes stay private.
An agent skill for writing and improving programs that call Jev.
SOURCE SCREENSHOTEXPAND ↗
from the source
git clone https://github.com/dbreunig/building-with-jev-skillcp -r building-with-jev-skill/skills/jev ~/.claude/skills/jev
Keep what you find. Notes stay private.
A web-search tool that uses Jev for query interpretation, source selection, and relevance ranking.
SOURCE SCREENSHOTEXPAND ↗
from the source
git clone https://github.com/superagents-lab/jev-search.gitcd jev-searchcorepack enablepnpm install --frozen-lockfilecp .dev.vars.example .dev.vars
Keep what you find. Notes stay private.
Callee lets repository-defined agent workflows run reusable Jev evaluations and carry the answers into later steps.
SOURCE SCREENSHOTEXPAND ↗
from the source
request := evaluationapi.Request{State: evidence, Questions: questions}evaluationResult, trace, err := service.Evaluate(ctx, config, request)
Keep what you find. Notes stay private.
Helps a Grok bot decide when to reuse work, stop retrying, limit research, delegate or bring in a person.
SOURCE SCREENSHOTEXPAND ↗
Keep what you find. Notes stay private.
A linter for mistakes that pass the type checker, like a comment promising null while the function throws.
SOURCE SCREENSHOTEXPAND ↗
Keep what you find. Notes stay private.
Let Jev check a Pi agent's proposed tool calls before they run.
SOURCE SCREENSHOTEXPAND ↗
noul
Is this relevant to the user's question?
Keep what you find. Notes stay private.
A Go MCP connector that gives AI agents direct access to Jev.
SOURCE SCREENSHOTEXPAND ↗
noul
Does this convey urgency?
Keep what you find. Notes stay private.
A Pi guardrail that steers agents away from risky, irrelevant, or low-quality tool use.
SOURCE SCREENSHOTEXPAND ↗
Keep what you find. Notes stay private.
Gives coding agents a Rust CLI and MCP server for deterministic SEO checks, live crawls, and optional Jev judgements.
SOURCE SCREENSHOTEXPAND ↗
Keep what you find. Notes stay private.
Runs semantic grep over a live source tree and returns scored matches with their file and line ranges.
SOURCE SCREENSHOTEXPAND ↗
Keep what you find. Notes stay private.
Run browser, computer-use, game, and robotics agents through one interface for Jev, Laya, and Cua-S1.
SOURCE SCREENSHOTEXPAND ↗
Keep what you find. Notes stay private.
Add plain-English rules to Oxlint and have Jev flag the code that appears to break them.
SOURCE SCREENSHOTEXPAND ↗
Keep what you find. Notes stay private.
Inspect Jev calls locally as they arrive, with requests, typed answers, timing and usage collected in a live visualizer.
SOURCE SCREENSHOTEXPAND ↗
Keep what you find. Notes stay private.
Find likely code files, log failures, or ticket labels with Jev before Claude spends context reading the survivors.
SOURCE SCREENSHOTEXPAND ↗
Keep what you find. Notes stay private.
Selects the project rules relevant to a Claude Code prompt or edited file, keeping unrelated instructions out of context.
SOURCE SCREENSHOTEXPAND ↗
Keep what you find. Notes stay private.
Find an agent skill that fits what you're doing now, with a Rust command-line tool that asks Jev to rank the options.
SOURCE SCREENSHOTEXPAND ↗
Keep what you find. Notes stay private.
Classifies batches of text before an agent reads them, exposing the filter through a portable plugin and MCP tool.
SOURCE SCREENSHOTEXPAND ↗
Keep what you find. Notes stay private.
CC Settings asks Jev whether a Claude Code prompt is broad enough to delegate.
SOURCE SCREENSHOTEXPAND ↗
from the source
const jev = key ? await jevNoul({ key, prompt, instructions: JEV_INSTRUCTIONS, timeoutMs: JEV_TIMEOUT_... : null;if (jev !== null) { if (jev < JEV_THRESHOLD) return;
Keep what you find. Notes stay private.
A Codex router that chooses the model, reasoning depth, and speed mode for each turn.
SOURCE SCREENSHOTEXPAND ↗
Keep what you find. Notes stay private.
Jeff reviews source files with a catalogue of Jev questions and reports pass, violation or inconclusive for each rule.
SOURCE SCREENSHOTEXPAND ↗
from the source
response, err := (*client).Evaluate(ctx, catalog.Model, state, misses)for code, answer := range response.Answers { answers[code] = answer store.Put(catalog.Model, state, misses[code], *answer.Noul)}
Keep what you find. Notes stay private.
A CLI and MCP server that searches repositories by meaning and returns exact excerpts with paths and line numbers.
SOURCE SCREENSHOTEXPAND ↗
Keep what you find. Notes stay private.
A Pi extension strips the full skills catalog and lets Jev rank which instructions should be loaded on demand.
SOURCE SCREENSHOTEXPAND ↗
Keep what you find. Notes stay private.
An agent skill that lets Hermes and other agents ask Jev for judgments.
SOURCE SCREENSHOTEXPAND ↗
from the source
mkdir -p "${HERMES_HOME:-$HOME/.hermes}/skills/autonomous-ai-agents/askjev"cp -R SKILL.md scripts "${HERMES_HOME:-$HOME/.hermes}/skills/autonomous-ai-agents/askjev...
Keep what you find. Notes stay private.
Recommends one installed Codex skill—or honestly none—from the catalog already available on the machine.
SOURCE SCREENSHOTEXPAND ↗
Keep what you find. Notes stay private.
A semantic Hono router that directs HTTP requests according to their meaning.
SOURCE SCREENSHOTEXPAND ↗
from the source
import type { JevResult } from 'hono-jev-router'const app = new Hono<{ Variables: { jev: JevResult } }>({ router: new JevRouter({ apiKey...app.on('jev', 'a request from an AI agent', (c) => { const { route, confidence, probabilities } = c.get('jev') // route: 'a request from an AI agent'
Keep what you find. Notes stay private.
A Solidity review heat map that points auditors toward the contracts and lines Jev considers most likely to hide bugs.
SOURCE SCREENSHOTEXPAND ↗
Keep what you find. Notes stay private.
Adds typed decision gates and review steps to a skills-based workflow for coding agents.
SOURCE SCREENSHOTEXPAND ↗
Keep what you find. Notes stay private.
Experiment with Jev from MoonBit through a typed client, command-line tools, games, policy gates, and reproducible reports.
SOURCE SCREENSHOTEXPAND ↗
Keep what you find. Notes stay private.
A Pi extension that turns four semantic risk checks into allow, warn, confirm or block for each tool call.
SOURCE SCREENSHOTEXPAND ↗
from the source
export const actionQuestions = { destructive: "Does executing the proposed action risk irreversible loss of user files/data, disca... data_leak: "Does this action send private local files, credentials or secrets to an external re...
Keep what you find. Notes stay private.
Screen agent tool calls and returned content with Jev, then review the judgments in shadow mode before enabling enforcement.
SOURCE SCREENSHOTEXPAND ↗
Keep what you find. Notes stay private.
Build programs and shell commands as chains of bounded Jev decisions through a TypeScript SDK and coding CLI.
SOURCE SCREENSHOTEXPAND ↗
Keep what you find. Notes stay private.
A Claude Code context filter that asks Jev whether each tool result should enter the conversation.
SOURCE SCREENSHOTEXPAND ↗
Keep what you find. Notes stay private.
Jev, briefly
A decision model, not a chatbot.
A chat model generates text token by token and hopes you parse it. Jev never generates a word. Your code sends the state of the world plus typed questions; Jev returns every answer in one parallel pass — typed values with calibrated probabilities and a confidence score your code can trust. About 70–500 ms end to end, $0.042 per million input tokens, output free. Trained with what TypeSafe calls Reinforcement Learning for Calibrated Decisions.
choice Pick one of these.
You define the options. Jev returns the chosen option, a probability for every option, and a confidence. The workhorse of routing, agents, and games — the answer space is always legal, so there is nothing to parse and nothing to hallucinate.
score Rate this on a rubric.
You describe ordered levels — say trivial / normal / critical. Jev returns the level, the probability of each, and a confidence. Scores turn fuzzy judgment (“how severe is this log line?”) into a number-free decision your code can branch on.
noul Is this statement true?
You assert a statement; Jev returns the probability that it is true — a noul. Moderation, verification, guardrails, “does this diff actually fix the bug?”: one question, one calibrated probability.
Official material lives at typesafe.ai and docs.typesafe.ai. JevMade is an independent community registry — not affiliated with or endorsed by TypeSafe AI.
Recurring lessons
Patterns that keep showing up.
One decision per tick
Games, drones, trading bots, and browser agents all converge on the same loop: serialize the state, ask one decisive question, act, repeat. Jev’s latency makes the loop feel instant — the model lives inside the control loop, not outside it.
Confidence as a gate
The answer says what; the confidence says whether to act. The most reliable entries threshold on confidence to route edge cases to a slower model or a human — automation with an honest escape hatch.
Shrink the choice space in code
Makers rarely ask Jev to pick from everything. Local tactics prune 225 gomoku moves to ~40; DOM filters turn a page into an element table; code narrows options, Jev judges within them.
Jev judges, LLMs talk
A recurring split: Jev makes every decision cheaply and instantly, and a small LLM is only invoked when a human-facing string must actually be written. Decision and generation are separate budgets.
Made something with Jev?
Send the link — repo, live demo, post, or video — plus a line on what it does and which primitives it uses. Every entry is verified against its primary source before it ships.
Submit your experimentSign in first so we can reply to you. Submissions are reviewed, not published automatically.