jevbooks

← All projects

jev-nlgrep

YehuiTang0316/jev-nlgrep

Natural-language grep for code, docs, logs, and text.

Scoring & Ranking95%Runner-up: Data Labeling & Curation

What it is

nlgrep is a developer tool that searches code, docs, logs, and text by meaning using natural-language queries, finding paraphrases and behavior descriptions even when exact words are forgotten.

How it uses Jev

nlgrep uses Jev for every match: it sends the query and included source content to Jev, which returns a probability that the condition is satisfied. Results are ranked by highest-scoring window and filtered by a configurable threshold.

Primitives:noul

Technique worth stealing

Use natural-language conditions with Jev Noul judgments to find code by meaning, with a probability threshold and ranked results.

Try it

npm install --global nlgrep; set JEV_KEY; run nlgrep "query" ./path

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-centricyes0.93
  • Shows a System One patternno0.21
  • Handles uncertaintyunclear0.53
  • Measuredno0.08
  • Runnableyes0.94
  • Worth recommendingunclear0.41
  • Model replicano0.09
  • Problem scopescore on a 0–2 scale1.82
  • About Jevyes0.98

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