jevbooks

← All projects

jevbrief

Parthkomalwad/jevbrief

Clean, traceable state briefings for TypeSafe's Jev model.

What it is

jevbrief sits between your data source and Jev, keeping only relevant facts, dropping noise with reasons, and asking one clear question. It is for developers using Jev who need traceable decisions.

How it uses Jev

Jev receives a briefed state and a typed question (e.g., which element to click, which log group explains an incident). It returns a probability distribution over options, a confidence, and a chosen answer, which jevbrief records in a trace.

Primitives:choice

Technique worth stealing

Rule-based filtering with reason codes for every dropped fact, plus a token budget to reduce input size.

Try it

pip install "jevbrief[all]" then run jevbrief ask with an adapter and goal.

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-centricyes0.62
  • Shows a System One patternno0.12
  • Handles uncertaintyno0.06
  • Measuredno0.13
  • Runnableyes0.90
  • Worth recommendingno0.35
  • Model replicano0.06
  • Problem scopescore on a 0–2 scale1.03
  • About Jevyes0.98

Signals by Jev jev-1.13.0, card written by DeepSeek V4.1 Flash from the README on 23 Sept 2026.