jevbooks

← All projects

jev-search

superagents-lab/jev-search · Homepage

Plain-language web search ranked by Jev relevance judgments.

Scoring & Ranking100%Runner-up: Classification & Routing

What it is

Jev Search lets users search the open web in plain language. Jev chooses sources, time ranges and search terms, then ranks results from Search1API. It returns links and snippets with visible relevance scores and editable filters, and no generated answers.

How it uses Jev

Jev answers typed questions about the user's request; the application uses those judgments to choose a query, sources and a time range. Jev then scores each result for relevance, and results are merged by URL, ordered by relevance, engine agreement and original rank, with lower-scoring results grouped separately.

Primitives:choicescorenoul

Technique worth stealing

Use Jev judgments to pick query, sources and time range, then score each result for relevance ranking.

Try it

Try the hosted demo at https://jev.s1.dev, or run locally with pnpm dev after setting API keys.

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-centricyes0.94
  • Shows a System One patternno0.36
  • Handles uncertaintyno0.05
  • Measuredno0.05
  • Runnableyes0.96
  • Worth recommendingunclear0.49
  • Model replicano0.05
  • Problem scopescore on a 0–2 scale0.84
  • About Jevyes0.98

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