Skip to content

jev-starter-examples

Two beginner Python apps using Jev: a message sorter and a support ticket router that send uncertain items to human review.

Projectprototype

Overview

Jev Starter Examples contains two small Python scripts. The message sorter asks whether a message is spam (Noul), urgent (Noul) and which category it belongs to (Choice), then files it into folders, marks urgent items and sends possible phishing to a review folder. The ticket router asks which team should handle a ticket (Choice), its urgency from 0 to 3 (Score), whether it is a security problem (Noul) and whether it contains enough detail (Noul), then writes priority queues and sends vague, risky or uncertain tickets to human review.

It is aimed at beginners. Both scripts run without an API key using fake keyword answers, and switch to real Jev when a key is set. The README shows real outputs, such as a bank sign-in alert scoring 50% spam and being routed to review, and a vague ticket picked as technical with 97% confidence but scored 6% on enough detail. A budget module records every call and refuses calls that would exceed a spending cap.

What the decision model decides

TypeQuestionAnswers
ChoiceWhich team should handle this ticket?—
ScoreHow urgent is this ticket (0-3)?0 · 1 · 2 · 3
Yes / noIs this a security problem?—
Yes / noDoes the ticket contain enough detail to act on?—

How it works

  1. Send all questions about an item in one request
  2. Thresholds at the top of each script map probabilities to actions
  3. Items between 40% and 70% spam, low-confidence or vague go to review with a reason
  4. Budget module logs spend and refuses calls over the cap

Strengths

  • Runs free without a key using fake answers
  • Shows how separate questions keep risks from being averaged away
  • Spending cap built in; README estimates about $0.0003 per run
  • CI tests

Caveats

  • Beginner examples, not a production system
  • Keyword fallback is not representative of Jev quality

Getting started

pip install -r requirements.txt && python sorter.py (add TYPESAFE_API_KEY to .env for real Jev).

At a glance

Type
Project
Status
prototype
Language
Python
Runtime
CLI
Has code
yes
Has tests
yes

Use cases and integrations

How curated lists describe it

Frequently asked questions

Can I try it without a TypeSafe key?

Yes, both scripts fall back to fake keyword-based answers.

Who makes the final decision?

Plain Python; Jev only returns probabilities and the scripts' thresholds decide.

What does a run cost?

The README estimates roughly $0.0003 per run of either script at $0.042 per million input tokens.

Related projects

  • Go CLI that asks Jev each message's category, urgency and frustration, sorts by urgency and lists low-confidence items for review.

    CLIprototypeGo
  • Python sandbox turning Jev issue-category and reproduction judgments into testable routing proposals, with offline replay.

    ProjectprototypePython
  • Data-incident router and use-case guide: registry facts first, then Jev Choice, Score and Noul with confidence-gated human review.

    ProjectprototypePython
  • Fifteen runnable Node recipes using Jev via OpenRouter for triage, dedupe, PII scanning, reranking and a browser agent.

    ArticlebetaJavaScript
  • Small Python demos of Jev: parallel Noul ticket checks with measured speedup, and a three-way Choice citation and claim checker.

    AppprototypePython
  • Worked Jev examples for ticket triage and reranking, runnable through OpenRouter without an early-access key, with offline tests.

    ProjectprototypePython