Jev Explained
Interactive playground that walks step by step through real Jev requests and their probabilities using your own key.
Overview
Jev Explained is an interactive web playground for learning how Jev makes typed, probabilistic decisions by running real requests. It explains the three primitives (Noul for binary questions, Choice for picking among known options, Score for ordered rubrics), that Jev reads the state once and answers all questions in parallel, and that confidence is a second axis: act automatically when it is high, ask a human when it is low.
The playground walks through four patterns with real requests, including an email spam classifier asking is_spam (Noul), folder (Choice) and suspicion (Score), and a stock buy-or-sell decision on market data. A live demo is hosted on Vercel, and users bring their own TypeSafe or Vercel AI Gateway key.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Yes / no | Is this email spam? | — |
| Choice | Which folder should this email go in? | — |
| Score | How suspicious is this email? | — |
How it works
- Open the playground and enter a key
- Pick an example pattern
- Send the state and typed questions
- Inspect each answer's probabilities and confidence
Strengths
- Hands-on learning with real requests
- Explains confidence-based routing clearly
- Hosted live demo
Caveats
- Requires your own key to run
- Educational scope
Getting started
Open https://jev-explained-repo.vercel.app/ and add a TypeSafe or Vercel AI Gateway key.At a glance
- Type
- App
- Status
- production
- Language
- TypeScript
- Runtime
- Web app on Vercel
- Has code
- yes
Use cases and integrations
How curated lists describe it
Interactive playground that walks through a typed request and its probabilities, with your own key.
MrJev/awesome-jev
Frequently asked questions
What will I learn?
How Noul, Choice and Score work and how to act on confidence.
Do I need a key?
Yes, a TypeSafe or Vercel AI Gateway key to run requests.
What examples are included?
Four patterns, including a spam classifier and a buy-or-sell decision.
Related projects
Read-aloud English tutor: streaming ASR transcribes, then Jev judges word by word whether each word was read correctly.
ProjectprototypeJavaScriptSix side-by-side demos where Jev makes the decision, Claude or Kimi writes the words, and code owns control flow.
AppprototypeTypeScriptChrome extension that scores each X timeline post with one batched Jev Noul request for engagement value, virality and tags such as spam.
Browser extensionprototypeJavaScriptChatGPT-style web demo where Jev answers by choosing from predefined or custom answer sets, such as Yes/No, Pirate or Tabloid.
AppprototypeTypeScriptSemantic mapping site that places phrases in 2D using Jev's scores for how strongly each relates to two chosen axis concepts.
Article- 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