jevbooks

← All projects

jselect

keltokhy/jselect

Source-linked evidence selection within a token budget using Jev Noul relevance.

Scoring & Ranking97%Runner-up: Data Labeling & Curation

What it is

jselect is a Python tool and package that selects relevant, source-linked passages from files, records, or agents, assembling a context within a token budget for AI agents. It is separate from jgrep.

How it uses Jev

jselect uses Jev Noul relevance judgments to score passages against a task. The result is used to select a diverse set of excerpts, which are assembled into a context with citations. Without credentials, it falls back to local lexical retrieval.

Primitives:noul

Technique worth stealing

Diversity-aware selection of source-linked evidence within a token budget.

Try it

uv tool install jev-select; jselect "task" data.jsonl --tokens 8000

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 patternno0.21
  • Handles uncertaintyno0.04
  • Measuredno0.06
  • Runnableyes0.91
  • Worth recommendingno0.31
  • Model replicaunclear0.43
  • Problem scopescore on a 0–2 scale1.88
  • About Jevyes0.90

Signals by Jev jev-1.13.0, card written by DeepSeek V4.1 Flash from the README on 21 Sept 2026.