jevbooks

← All projects

jev-axi

jaykang-heo/jev-axi

Agent-ergonomic CLI for TypeSafe Jev typed judgments

Agent Decisions74%Runner-up: Evaluation & BenchmarkingMany questions, one request

What it is

A TypeScript CLI that asks TypeSafe's Jev for typed judgments about a text: pick one option from a closed set, give a yes/no probability, or score against a rubric. Built for agents and developers who threshold on numbers.

How it uses Jev

Each command sends a state (text or JSON) plus a typed question to Jev. choice returns option probabilities, noul a single 0-1 probability, score a probability-weighted level. Results are evidence to threshold on in your own code, not decisions.

Primitives:choicenoulscore

Technique worth stealing

Threshold on both winning_probability and confidence; a noul's number is not on the same scale as a choice's confidence.

Try it

npm install -g github:jaykang-heo/jev-axi#jev-axi-v0.1.0, then jev-axi

View on GitHub

judged by Jevjev-1.13.0

Evidence

Each line is one question put to Jev about the README. ≥ 0.60 reads as yes, ≤ 0.40 as no; in between Jev is not making a call.

  • Jev-centricyes0.96
  • Shows a System One patternno0.30
  • Handles uncertaintyno0.12
  • Measuredno0.04
  • Runnableyes0.98
  • Worth recommendingunclear0.41
  • Model replicano0.04
  • Problem scopescore on a 0–2 scale1.01
  • About Jevyes0.98

Signals by Jev jev-1.13.0, card written by DeepSeek V4.1 Flash from the README on 20 Sept 2026.