yang3kc/daily_arxiv_digest
19 stars · Last commit 2026-07-23
Daily digest of interesting arXiv papers — an unattended LLM pipeline for cron, plus a zero-dependency agent skill your AI agent runs itself
README preview
# Introduction Automatically generate a daily digest of interesting arXiv papers. Having trouble catching up with the new arXiv papers everyday? This tool fetches the latest arXiv papers, uses an LLM to rate their relevance to the topics of your interest, and writes a daily digest — a human-readable markdown file plus a machine-readable JSON file — so it can run unattended in a cron job or be driven by another agent. It comes in two forms: - **Daily digest pipeline** (this section and below) — the full pipeline with its own LLM calls, designed to run unattended (e.g. daily cron). Needs an API key. - **[Agent skill](#agent-skill)** — a standalone skill for AI agents (e.g. Claude Code): the agent fetches the papers and judges relevance itself. No API key, no dependencies, no repo clone needed — just any Python 3. # How to use ## API key The tool supports three LLM providers, selected via `llm_provider` in `config.json`. Export the API key for the provider you use: | Provider | Environment variable |