Search arXiv in One Command — No API Key, No Tokens

Keeping up with AI research is exhausting. New papers drop daily. Most "paper discovery" tools require an account, burn API tokens on every search, or give you a bloated UI when all you wanted was a list.

Here's what I use instead:

npx -y @taprun/cli arxiv search --keyword "LLM" \
  | npx -y @taprun/cli sort --field published \
  | npx -y @taprun/cli table

Output: 20 papers sorted newest-first, with title, authors, published date, abstract, and URL — in under 2 seconds.

No account. No API key. No AI tokens consumed. First run downloads a ~30MB binary and caches it; every subsequent call is instant.

Why This Works

The arXiv Atom API has been public and stable for 15 years. arxiv/search is a Tap skill — a 20-line deterministic program that calls it directly. AI wrote it once. It runs forever at $0.

The Unix Pipeline Model

Every Tap skill is a composable Unix filter. Data flows as JSON:

# Search only
npx -y @taprun/cli arxiv search --keyword "RAG"

# Search → sort by date → filter recent → display
npx -y @taprun/cli arxiv search --keyword "RAG" \
  | npx -y @taprun/cli sort --field published \
  | npx -y @taprun/cli filter --field published --gt "2025-01-01" \
  | npx -y @taprun/cli table

Each command reads JSON from stdin, writes JSON to stdout. Exactly how Unix tools work.

Use It in CI

# GitHub Actions — daily paper digest
- name: Paper digest
  run: |
    npx -y @taprun/cli arxiv search --keyword "LLM agents" \
      | npx -y @taprun/cli sort --field published \
      | npx -y @taprun/cli limit --n 5 \
      > papers.json

200+ Skills, Same Pattern

arXiv is one of 200+ community skills that follow the same pattern: call an API, return structured rows, compose with any other skill.

Skill	Returns
arxiv/search --keyword X	Papers matching keyword
github/trending	Trending repos today
reddit/search --keyword X	Posts matching keyword
stackoverflow/hot	Hot questions

Browse and contribute: github.com/LeonTing1010/tap-skills

---

Try it now — no install required, works on any machine with Node.js:

npx -y @taprun/cli arxiv search --keyword "your topic" \
  | npx -y @taprun/cli sort --field published \
  | npx -y @taprun/cli table

→ taprun.dev