Skip to content
JevDirectory.orgTraffic count unavailable. View site traffic

Jev directory

CommunityJev Repos & SDKs

Jev Harness — AntonioCoppe

A TypeScript library that wraps Jev answers in a policy, a confidence gate, shadow mode, reusable recipes, and an offline eval CLI, so structured decisions can drive actions an application ships.
See highlightsOpen github.com
Stars
15
Language
TypeScript
License
MIT
Reading time
4 min
Published by
Community
Author
AntonioCoppe
Last verified
Sep 2026

About this project

A small TypeScript library that turns TypeSafe Jev answers into actions you can ship. Jev returns structured decisions (choice / score / noul) with confidence scores. This harness adds the missing pieces around that API: a policy, a confidence gate, shadow mode, reusable recipes, and an offline eval CLI. It is not affiliated with TypeSafe. You need a TypeSafe API key to call Jev. Calling Jev once is easy. Running it in production usually needs more:

Quoted from the project's README. Read the full README.

Tagscommunitytypescriptagentsguardrailsconfidence
Works forJev for EvaluationJev for VerificationJev for Decision gating
Used inJev for Software & DevTools

Highlights

  • Measured: filtering 24 people rows with a natural-language predicate took 1.3 s with Jev and the harness versus 48.9 s for the Claude Code CLI with tools off.
  • Policy maps answers to actions such as notify, suppress, place, or skip, and low confidence routes to review, escalate, or suppress instead of guessing.
  • Fixtures assert on the action and an optional confidence band, so policy changes can be tested offline with npm run eval.
  • Seven recipes cover agent handoffs, alert gating, row filters, high-frequency allow/deny, prediction markets, sports bets, and output verification.
  • Shadow mode logs what the policy would do without changing live behavior: result.action becomes shadow_noop while intendedAction holds the choice.

Quickstart

ts
import { DecisionHarness, choice, noul } from "jev-harness";

const result = await new DecisionHarness().run({
  id: "alert-42",
  state: { title: "Disk 92% on db-3", service: "payments" },
  questions: {
    disposition: choice("What should we do?", { notify: "Page someone now", suppress: "Ignore as noise" }),
    needs_human: noul("Does a human need to look at this?"),
  },
  policy: { minConfidence: 0.55, onLowConfidence: "review", decide: ({ answers }) => (answers.needs_human.noul >= 0.6 ? "notify" : "suppress") },
});
console.log(result.action, result.confidence, result.reason);

Watch out

MIT-licensed and not affiliated with TypeSafe; needs Node 20+ and a TYPESAFE_API_KEY, and the README explicitly says it is not a /compact replacement for transcript compaction.

Questions about this entry

Short answers drawn from the linked project or page.

Reactions & coverage

Posts, threads, and videos about this entry from around the web.

X: Agent Handoff Gate: agents verify what they hand off

X: Jev as an LLM guardrail inside Agentgateway

X: Vercel's fx safety reviewer, measured at 18× faster

Reddit: Early experiments wiring Jev into Pi as a safety layer and model router.

Other listings of this project

Forks and copies of this project in the directory. This entry is the canonical one.

Jev Harness
9GitHub stars
An agent harness where Pi runs the loop, Gemini does the work, and Jev makes the small decisions: which model tier to use, which runbook sections to load, whether a tool call may run on a real Postgres database, and whether the report holds up.
Jev Agents & MCP#community#python#agents
CommunityPromtEngineer/jev-harnessPython
Jev Harness
6GitHub stars
A minimal coding agent that filters every tool result through Jev before the main model reads it, plus the A/B harness that measured pass@1 and cost. The final TB4 follow-up closed the experiment without an end-to-end win.
Jev Benchmarks & Evaluations#community#catalog-import#madewithjev
Communityastro-han/jev-harnessPython
jev-harness
12GitHub stars
Decision layer for coding agents that triages test failures deterministically before spending frontier tokens. Offline heuristics handle missing packages, flaky transients, and loop breaking, with optional Jev micro-decisions and CLI, MCP, hook, and SDK use.
Jev Demos & Experiments#community#awesomejev#catalog-import
Communityismaelsoilet/jev-harnessPython
jev-harness
30GitHub stars
A research-stage proposal-review contract in TypeScript: an LLM proposes one action, Jev answers four pinned questions, and code produces evidence the host can consider.
Jev Tools & CLIs#community#typescript#agents
CommunityTypeSafeAI/jev-harnessTypeScript

From the Jev guides

Original write-ups on what this entry is for.

From the Jev AI model community

Posts from builders shipping with Jev right now.

Follow @typesafeai