Open Source Tool Automatically Filters arXiv Papers
usedtobreath · reddit · 2026-07-13
The author built an open-source tool called Research Radar to solve the problem of "only a few out of hundreds of daily papers being truly relevant."
What it does
- Crawls new papers from your followed arXiv categories
- Uses a markdown file to describe your research interests and scores abstract relevance
- Downloads PDFs for high-scoring papers, extracts the full text, and uses a powerful model to generate summaries, key insights, limitations, and relevance to your research
- Outputs a daily HTML digest, with optional Telegram alerts
Design Features
- Research domain info is stored in a single markdown file, making it easy to switch fields
- Only the "filtering" and "deep reading" steps call models; the rest of the workflow uses deterministic Python scripts
- The backend supports Claude Code / Codex CLI, OpenAI-compatible APIs, and even local Ollama / vLLM
- The author provides rough benchmarks for tokens, cost, latency, and quality
The author notes they've used it continuously for a month and found it genuinely useful, though it hasn't been fully validated in other domains yet, and welcomes feedback.
More from coding & agent
- A better path to agent autonomy is running waves, finding friction, and iterating — JnBrymn · 2026-07-22
- Coding agents are heading toward an AI-writes, AI-reviews, human-approves workflow — aftahi_ai · 2026-07-22
- oMLX 0.5.2 adds Mac menu-bar stats, low-bit decode kernels, and faster downloads — awnihannun · 2026-07-22
- GitHub review bot hits its PR limit and forces a 39-minute cooldown — DanielLockyer · 2026-07-22
- Max reasoning effort appears to be mobile-only in Codex Remote, not desktop — GabGarrett · 2026-07-22
- A Reddit demo argues online stores should expose carts and pricing through MCP — gelembjuk · 2026-07-22