Skip to content
Metatext Decide — typed decisions for your agents, with a probability on every answerGet early access
Metatext
Sign Up

Jev Showcase

Unofficial Jev operator lab: offline-first FastAPI app with primitives playground, cost lab, pattern pages and a TypeScript harness.

Agent skillbetaMIT
View on GitHub ↗

Overview

Jev Showcase by cobusgreyling is an unofficial companion for TypeSafe's Jev that focuses on details the author thinks most posts miss. Its summary table covers points such as: Jev is not a chatbot and should be paired with a generative model; it has three primitives, Choice, Score and Noul; many questions can share one state in one call, which the README puts at about 12 times cheaper than serial calls; output tokens are free; Noul has no confidence field, so its thresholds should not be copied onto Choice; there is no fine-tuning; and counting, dates and arithmetic should stay in code.

The repository is an offline-first FastAPI lab with operator cards, a cost lab, a primitives playground and pattern pages, plus a smart-home fan-out demo and an optional live judge against TypeSafe. A gallery of four agent skills puts the primitives next to parallel fan-out, a router and a guardrail, and a TypeScript harness CLI packages the same questions for agent loops covering model routing, semantic retrieval, guardrails and reasoning-trace classification. The README cites jev-1.13.0, released 15 September 2026.

What the decision model decides

TypeQuestionAnswers
ChoiceWhich option fits this state? (playground example)(inferred)—
Yes / noShould this device action run? (smart-home fan-out)(inferred)yes · no

How it works

  1. Clone and install requirements
  2. Run python app.py
  3. Explore the operator lab and pattern pages offline
  4. Optionally enable the live judge with a key
  5. Use the harness CLI in agent loops

Strengths

  • Works offline for learning
  • Collects non-obvious API details
  • Runnable patterns in one codebase

Caveats

  • Unofficial; check vendor docs for current behavior
  • Many facts are vendor figures
  • Live features need a TypeSafe key

Getting started

git clone https://github.com/cobusgreyling/Jev.git, pip install -r requirements.txt, python app.py, then open http://127.0.0.1:7872

At a glance

Type
Agent skill
Status
beta
Language
Python
Runtime
FastAPI web app (plus TypeScript CLI)
Has code
yes
License
MIT
Stars
137

Use cases and integrations

How curated lists describe it

  • Pattern gallery: a runnable app with four agent skills that put `Choice`, `Score` and `Noul` next to parallel fan-out, a router and a guardrail in one codebase.

    yibie/awesome-jev

  • Unofficial TypeSafe Jev showcase — System One decisions, not chat.

    logicrw/awesome-jev-projects

  • Unofficial TypeSafe Jev showcase — System One decisions, not chat.

    wh000wh000/awesome-jev-live

  • Unofficial TypeSafe Jev showcase — System One decisions, not chat.

    hellogumbo/awesome-jev

  • Offline-first FastAPI lab for learning Jev: operator cards, a cost lab, a primitives playground and pattern pages, with an optional live judge against TypeSafe.

    MrJev/awesome-jev

Frequently asked questions

What is it for?

Learning how to build with Jev through runnable patterns, a cost lab and a primitives playground.

Do I need an API key?

Not for the offline lab; the live judge needs one.

Is it official?

No. It describes itself as an unofficial companion.

Related projects

  • Building with Jev★ 155

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

    Agent skillproduction
  • jev-agent-toolkit

    Portable Agent Skill teaching coding agents to use Jev as a bounded judgment layer, plus an optional MCP bridge.

    MCP serverbetaTypeScript
  • jev-anything

    Independent agent skill that makes a coding agent design, build, test and tune a bounded Jev decision layer with a contract first.

    Agent skillbeta
  • jev-architect

    Agent skill that guides you in finding, designing and testing Jev decision loops, including when to keep logic in code.

    Agent skillbeta
  • jev-mode (tiffygk)

    Claude Code and Codex skills plus a study course for building with Jev, including Jevaluate, which rates how well projects use Jev.

    Projectbeta
  • jev-skill (marcodicesare-dev)

    Community Jev skill for Claude Code and Codex: a guide, a dependency-free Python CLI and 5 measured question sets.

    Agent skillbetaPython