What it is
A configurable Python CLI for researchers: it fetches recent arXiv papers, evaluates how well each matches a JSON research profile, and writes daily Markdown and JSON relevance rankings.
How it uses Jev
For each paper, the research profile, title, abstract, and categories are sent to TypeSafe AI, which answers the relevance question using P(True). The returned probability is compared against configurable thresholds to rank, filter, and sort papers in the reports.
Primitives:score
Technique worth stealing
Define the research profile entirely in JSON, including negative topics and evaluation guidance, so no research domain is hard-coded.
Try it
gh repo clone LYchoon/paper-radar-jev, uv sync, set TYPESAFE_API_KEY, then uv run paper-radar --config config/config.json.
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-centricno0.35
- Shows a System One patternno0.19
- Handles uncertaintyno0.08
- Measuredno0.05
- Runnableyes0.97
- Worth recommendingno0.17
- Model replicano0.05
- Problem scopescore on a 0–2 scale1.02
- About Jevyes0.65
Signals by Jev jev-1.13.0, card written by DeepSeek V4.1 Flash from the README on 20 Sept 2026.