Academic Paper Search API: OpenAlex, Crossref & arXiv

Search 250M+ scholarly works and look up DOIs or arXiv ids in one call. Results from OpenAlex, Crossref and (optionally) arXiv are merged into one deduplicated record per paper: title, DOI, arXiv id, year, venue, authors, institutions, abstract, citation count, open-access status, PDF link, licence, topics and retraction flag. Built for literature reviews, research agents and RAG. $1 per 1,000 papers. Empty searches and unknown ids are free.

What does it do?

• Sources: OpenAlex and Crossref by default. arXiv is opt-in and off by default: add arxiv to sources to also query arXiv's own API (slower, 1 request every 3 s). arXiv papers are still found through OpenAlex without it.
• Keyword search across OpenAlex and Crossref (and arXiv if enabled), with year range, open-access-only and sort (relevance, most cited, newest).
• DOI lookup in bulk (up to 10,000 per run) and arXiv id lookup (1706.03762, 2005.11401v4).
• Deduplication and merge: the same paper from several sources becomes one row (matched by DOI, arXiv id or title + year); fields are combined and sources tells you where it came from. You pay once.
• Abstracts rebuilt from OpenAlex's inverted index, Crossref JATS or arXiv summaries.
• Two views: Papers and Abstracts.

Input

Field	Default	What it does
queries	–	Keyword searches (max 100).
dois	–	DOIs to look up (10.1038/nature14539 or https://doi.org/...).
arxivIds	–	arXiv ids to look up.
sources	openalex, crossref	Add arxiv for arXiv's own search (slower: 1 request / 3 s).
maxResultsPerQuery	25	Per query and source (max 1,000); duplicates merged.
yearFrom / yearTo	–	Publication year range (searches only).
openAccessOnly	false	Only free-to-read papers.
sort	relevance	relevance, cited, newest.
includeAbstract	true	Return abstracts.
includeAuthorNames	true	Return author names (off = only authorCount).
contactEmail	–	Sent as mailto for Crossref/OpenAlex polite pools (recommended).
openalexApiKey	–	Your free OpenAlex key (secret) for your own daily budget.

{
  "queries": ["retrieval augmented generation", "protein structure prediction"],
  "dois": ["10.1038/nature14539"],
  "arxivIds": ["1706.03762"],
  "maxResultsPerQuery": 20,
  "yearFrom": 2020,
  "sort": "cited"
}

Output

One row per paper (shortened):

{
  "title": "Attention Is All You Need",
  "arxivId": "1706.03762",
  "openalexId": "W2626778328",
  "type": "preprint",
  "authors": ["Ashish Vaswani", "Noam Shazeer", "Niki Parmar", "..."],
  "authorCount": 8,
  "institutions": ["Google (United States)", "University of Southern California", "University of Toronto"],
  "abstract": "The dominant sequence transduction models are based on complex recurrent or convolutional neural networks...",
  "citedByCount": 78342,
  "isOpenAccess": true,
  "pdfUrl": "https://arxiv.org/pdf/1706.03762",
  "topics": ["Natural Language Processing Techniques", "Topic Modeling"],
  "keywords": ["transformer models", "attention mechanism", "machine translation"],
  "isRetracted": false,
  "url": "https://arxiv.org/abs/1706.03762",
  "sources": ["openalex"],
  "lookup": "arxiv:1706.03762"
}

Queries or ids without results are listed with error / errorCode (NO_RESULTS, NOT_FOUND, INVALID_DOI, ALL_SOURCES_FAILED) and are not charged. The SUMMARY record shows per-source errors (for example an OpenAlex rate limit).

How much does it cost?

Event	Price
Paper (unique, merged)	$0.001

• 1,000 papers ≈ $1. A 50-paper literature scan ≈ $0.05.
• Apify's tiny run-start fee applies (512 MB default). Set Maximum cost per run to cap spending.

Sources, licences and limits

• OpenAlex metadata is CC0. OpenAlex gives keyless users a small shared daily budget; for regular or large runs add your free openalexApiKey. If OpenAlex is rate-limited, Crossref still answers and the SUMMARY reports it.
• Crossref metadata is free to reuse; abstracts may be copyrighted by the publisher: fine for search, screening and analysis, check before republishing.
• arXiv metadata is CC0 ("Thank you to arXiv for use of its open access interoperability"). The Actor follows the arXiv API terms (one request every 3 seconds), links to PDFs but never downloads or serves them, and is off by default: it is only queried when you add arxiv to sources (arXiv papers are also found via OpenAlex).
• Not affiliated with or endorsed by OpenAlex/OurResearch, Crossref or arXiv/Cornell University.

Personal data

Author names are public bibliographic data and are included by default, because citations need them; switch off includeAuthorNames to get only authorCount. ORCID iDs, e-mail addresses and raw affiliation strings (which can contain e-mails) are never returned; only institution names are kept.

Use it from code or an AI agent

curl -X POST "https://api.apify.com/v2/acts/thriving_adder~academic-research-papers/run-sync-get-dataset-items" \
  -H "Authorization: Bearer YOUR_APIFY_TOKEN" -H 'Content-Type: application/json' \
  -d '{"queries":["large language model evaluation"],"maxResultsPerQuery":10}'

Also available as a tool in Apify's MCP server, so research agents can search and cite papers.

Use cases

• Literature reviews: pull the most cited or newest papers on a topic, with abstracts, into a spreadsheet for screening.
• RAG and research agents: search and cite papers by DOI, with open-access PDF links when they exist.
• Reference cleanup: look up a list of DOIs or arXiv ids and get consistent metadata.
• Research monitoring: schedule weekly searches for new papers on your topics.
• Bibliometrics: citation counts, venues, topics and open-access status for a set of papers.

FAQ

Why fewer rows than maxResultsPerQuery × sources? Duplicates across sources are merged into one row (and charged once).

Full text? Not included; use pdfUrl / oaUrl to open the open-access copy.

Problems or ideas? Open an issue in the Issues tab. We answer within 2 business days.

Related Actors

By the same publisher, built on the same safe, pay-per-result design:

• Web Page to Markdown & llms.txt Generator: convert open-access pages and docs to Markdown for RAG.
• SEC EDGAR Filings & XBRL Financials: company filings and financial data.
• Malaysia & Singapore Open Data API: official Malaysian and Singapore statistics.