arXiv — Scientific Papers & Preprints Search to JSON

Search millions of arXiv preprints — as clean JSON. Pass search queries and get back a tidy row per paper: title, authors, abstract, categories, DOI, journal reference, submission dates and PDF URL. Powered by the official arXiv API — no API key.

The easy way to build research and preprint datasets for science, RAG and monitoring.

Who uses this

• 🎓 Researchers & academics — track new preprints in your field.
• 🤖 AI / LLM & RAG pipelines — ingest abstracts and metadata at scale.
• 📊 Research analysts — study topics, authors and categories.
• 🏢 R&D & competitive intel — monitor a field's latest work.
• 🧑‍💻 Developers — a clean arXiv search endpoint by query.

What you get — one clean row per paper

Field	Description
title	Paper title
authors + authorCount	Authors
summary	Abstract
primaryCategory + categories	arXiv categories
published + updated	Submission dates
doi	DOI (if available)
journalRef	Journal reference
comment	Author comment (pages, figures)
absUrl + pdfUrl	Abstract page & PDF
id, query, scrapedAt	Provenance

Example

Input

{ "queries": ["all:large language models", "cat:cs.LG AND ti:transformer"], "maxResultsPerQuery": 30, "sortBy": "relevance", "sortOrder": "descending" }

Output (one item)

{
  "id": "1706.03762v7",
  "title": "Attention Is All You Need",
  "authors": ["Ashish Vaswani", "Noam Shazeer"],
  "primaryCategory": "cs.CL", "categories": ["cs.CL", "cs.LG"],
  "published": "2017-06-12T00:00:00Z",
  "absUrl": "https://arxiv.org/abs/1706.03762",
  "pdfUrl": "https://arxiv.org/pdf/1706.03762"
}

Why this actor

• ✅ Official source — the arXiv API, always current.
• 🔎 Powerful query syntax — all:, ti:, au:, abs:, cat: with AND/OR.
• 📝 Full abstracts — ready for LLM/RAG ingestion.
• 🔀 Sort options — by relevance, submission or last update.
• ⚡ Fast & affordable — pay only per paper returned. No API key.

Input options

• queries (required) — arXiv search queries (use field prefixes like cat:cs.AI).
• maxResultsPerQuery — cap papers per query.
• sortBy — relevance, lastUpdatedDate or submittedDate.
• sortOrder — descending or ascending.

FAQ

How do I search a category? Use cat: — e.g. cat:cs.LG for machine learning, cat:hep-th for high-energy physics.

Can I combine terms? Yes — cat:cs.CL AND ti:translation, au:Bengio, etc.

Do I get the PDF link? Yes — pdfUrl and absUrl are included.

How fresh is the data? Live from the arXiv API on every run (with a retry for transient hiccups).

Can I export to CSV/Excel/Google Sheets? Yes — the dataset exports to JSON, CSV, Excel and HTML, or pull it via the API.

What does it cost? Pay-per-paper — you're only charged for papers actually returned.

---

Tip: sort by submittedDate on your field's category to build a daily feed of the newest preprints.