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

JEV

Colab notebook experiment that analyzes a movie review with one Noul, one Choice and one Score question to Jev.

Projectprototype
View on GitHub ↗

Overview

This JEV repository by sihaam20 is a small learning experiment: a Google Colab notebook that sends a movie review to Jev with three questions. A Noul asks whether the review is positive, a Choice picks positive, negative or mixed, and a Score rates sentiment strength. Python logic then acts on the result, for example treating a review as positive when the probability is above 0.8 and sending anything else to human review.

The README walks through installing requests, storing JEV_API_KEY in Colab secrets, loading it, and running the review. Its example review returns positive at 0.99, category positive and strength extremely positive. It lists ticket classification, support routing, document classification, workflow automation and agents as possible uses, though none are implemented. The repository contains only the notebook, a README and a requirements file, so it is best read as a first-contact tutorial rather than a tool.

What the decision model decides

TypeQuestionAnswers
Yes / noIs the review positive?yes · no
ChoiceIs the review positive, negative or mixed?Positive · Negative · Mixed
ScoreHow strong is the sentiment?—

How it works

  1. Add JEV_API_KEY to Colab secrets
  2. Load the key in the notebook
  3. Send a review with three questions
  4. Branch on the positive probability

Strengths

  • Shows all three primitives in one call
  • Simple human-review fallback
  • Easy to run in Colab

Caveats

  • Single example, no evaluation
  • Notebook only
  • No license stated

Getting started

Open jev.ipynb in Google Colab, add JEV_API_KEY as a secret and run the cells

At a glance

Type
Project
Status
prototype
Language
Python
Runtime
Jupyter notebook (Colab)
Has code
yes
Has tests
no

Use cases and integrations

How curated lists describe it

Frequently asked questions

What does Jev decide?

Whether a movie review is positive, which sentiment category it fits, and how strong it is.

What happens to uncertain reviews?

Below 0.8 probability of positive, the example sends it to human review.

Is it production-ready?

No. It is a small learning experiment.

Related projects

  • jev-lab

    Zero-dependency single-page Jev classifier demo showing request JSON, probabilities, confidence, latency and token usage.

    AppprototypePython
  • JevTools

    Local Python toolkit and web dashboard for trying Jev via OpenRouter or TypeSafe, with review analysis, topic routing and request history.

    ServiceprototypePython
  • dsh-jev

    DeepSeek Harness plugin that classifies each user message's emotion, urgency and department with Jev and injects it as context.

    PluginbetaTypeScript
  • JevBystander

    Zero-dependency Android app that judges the visible WeChat message's intent, emotion, urgency and reply posture with Jev.

    AppbetaJava
  • JevGuide

    Android WeChat assistant where Jev judges the chat and scores the relationship, and a chat model drafts three replies.

    AppbetaKotlin
  • JevIntent

    WeChat plugin that classifies a long-pressed message's intent, emotion, urgency and suggested reply posture with Jev.

    PluginbetaJava