Latest N is not a complete daily window
The README explicitly warns that papers older than the fetched latest-N set can be missed after a long gap. A “daily” report name does not guarantee exhaustive daily coverage.
Inspect the supporting source ↗REPOSITORY ANALYSIS
paper-radar-jev
An automated research paper radar that fetches the latest papers from arXiv, evaluates their relevance to a configurable research profile using TypeSafe AI, and ranks them by relevance score. Designed for personalized, daily literature discovery across different research domains.
E2 · source reviewedPublic material or source evidence was reviewed. A fixed-version source review requires checking a pinned revision; this is not a reproduction, performance endorsement, or safety guarantee.
- Stars
- 1
- Forks
- 0
- License
- MIT
- Last commit
- 2026-09-20
What Jev does here
Per paper, Jev answers the overall relevance Noul and configured dimensions together; code applies display thresholds while retaining every successful evaluation.
Separates a short reading list from the complete evaluation archive, with failed items remaining retryable.
Fixed-version documentation and implementation material were inspected on 2026-10-24. The paper-fetch and live evaluation flow was not run here. The review relies on the pinned protocol documentation and model definitions, not a relevance benchmark.
What task does this project fit?
Build a research-paper shortlist without silently losing failed or low-scoring evaluations.
Material analysed：
From input to action
- Fetch the latest configured number of arXiv papers, normalizing versioned IDs.
- Evaluate each paper against the research profile and optional dimensions.
- Persist successful evaluations before rendering filtered Markdown and complete JSON reports.
Implementation details that matter
Display filters differ from stored evidence
Markdown applies relevance and display limits; JSON retains all successfully scored papers. API failures remain retryable instead of being marked seen.
Inspect the supporting source ↗When to choose it
Choose it for a bounded personal reading queue. Raise retrieval coverage or use a different feed strategy when missing any paper is unacceptable.
How to inspect it yourself
Inspect --check-config and the Output and recovery section. Confirm that the command validates configuration only, and compare what belongs in Markdown, JSONL and the seen file after a failed evaluation.
Limits of this review
- The paper-fetch and live evaluation flow was not run here.
- The review relies on the pinned protocol documentation and model definitions, not a relevance benchmark.