Academic Paper Scraper: arXiv, PubMed, Crossref & OpenAlex API

Search academic papers across Crossref, arXiv and PubMed (plus OpenAlex with your free key) in one run, merged and de-duplicated by DOI, arXiv ID, PubMed ID and title. You get one clean record per paper: title, authors, year, journal, abstract, DOI/PMID/arXiv ID, citation count, open-access status, license and PDF link.

• 🔓 A Google Scholar alternative that doesn't scrape. Every result comes from an official open scholarly API, so there are no CAPTCHAs, no proxies and no breakage when a page layout changes.
• 🧬 Merge and de-duplicate across databases, with a relevance guard that drops loose matches.
• 🔔 Research alerts: newest-first + only-new-papers + a schedule.
• 💲 $2.00 per 1,000 papers, no start fee, no monthly rental. Apify compute is included.

What does this academic paper scraper do?

For each search query it calls each selected source's official API, normalizes the results into one schema, merges records that are the same paper, applies your filters, and saves one row per unique paper.

In a test run on the Apify platform on 2026-10-27, the default query CRISPR base editing returned 115 raw results (Crossref 50, arXiv 15, PubMed 50) and saved 85 unique, relevant papers in under 4 seconds.

Why use it?

• 🔎 4 scholarly databases in one Actor:
  • Crossref: journals, books and proceedings, 180M+ DOIs;
  • arXiv: preprints in physics, CS, math, quantitative biology and more;
  • PubMed: biomedical literature, with MeSH terms;
  • OpenAlex: 250M+ works with citation counts and open-access links (add your free key).
• 🧬 One record per paper. The same paper found in Crossref and PubMed becomes one record listing both sources, with the best fields from each.
• 🎯 Relevance guard. Requires every query word (stemmed) or quoted phrase to appear in the title, abstract or subjects. You can turn it off.
• 📅 Filters: year range, open access only, abstract required, minimum citations, excluded title words.
• 🧾 Structured identifiers (DOI, PMID, PMCID, arXiv ID, OpenAlex ID), ready for reference managers, spreadsheets, literature-review pipelines and RAG/LLM apps.
• 🔌 Works with the Apify API, webhooks, Zapier, Make, n8n and AI agents through the Apify MCP server.

Use cases

• Literature reviews and systematic-review screening: pull candidates from several databases into one de-duplicated list.
• Research alerts: new papers on your topic every week, sent to Slack, email or Notion.
• AI and RAG pipelines: feed titles, abstracts and open-access PDF links into embeddings or agents.
• Competitive and technology intelligence: track new publications on a technology, compound or method.
• Bibliometrics: citation counts, venues, years and open-access share for a topic.

How to use it

1. Click Try for free.
2. Enter one or more search queries, e.g. CRISPR base editing or "retrieval augmented generation". Quotes mark exact phrases.
3. Pick sources and results per source, and optionally a year range or open access only.
4. Click Start and download JSON, CSV or Excel, or use the API.
5. For alerts: set Sort by to Newest first, turn on Only papers not seen in previous runs, and add a Schedule.

Input example

{
  "queries": ["CRISPR base editing", "\"prime editing\" delivery"],
  "sources": ["crossref", "arxiv", "pubmed"],
  "maxResultsPerSource": 50,
  "yearFrom": 2023,
  "openAccessOnly": false
}

Output example

A real record from the test run on 2026-10-27, found in both Crossref and PubMed and merged into one row:

{
  "id": "doi:10.1089/crispr.2020.29101.dwa",
  "title": "CRISPR-Free Mitochondrial DNA Base Editing",
  "authors": ["Douglas C. Wallace"],
  "authorsText": "Douglas C. Wallace",
  "year": 2020,
  "publicationDate": "2020-08-01",
  "venue": "The CRISPR Journal",
  "publisher": "SAGE Publications",
  "type": "journal-article",
  "abstract": null,
  "doi": "10.1089/crispr.2020.29101.dwa",
  "arxivId": null,
  "pmid": "32833534",
  "pmcid": null,
  "openalexId": null,
  "url": "https://doi.org/10.1089/crispr.2020.29101.dwa",
  "pdfUrl": "https://journals.sagepub.com/doi/pdf/10.1089/crispr.2020.29101.dwa",
  "isOpenAccess": null,
  "license": "https://journals.sagepub.com/page/policies/text-and-data-mining-license",
  "citationCount": 1,
  "subjects": ["CRISPR-Cas Systems", "DNA, Mitochondrial", "Gene Editing", "Humans", "Mitochondrial Diseases"],
  "language": "eng",
  "sources": ["crossref", "pubmed"],
  "sourceUrls": {
    "pubmed": "https://pubmed.ncbi.nlm.nih.gov/32833534/",
    "crossref": "https://api.crossref.org/works/10.1089/crispr.2020.29101.dwa"
  },
  "query": "CRISPR base editing",
  "retrievedAt": "2026-10-27T14:44:52.934097Z"
}

A field the sources don't provide is null (this editorial has no abstract in either source; subjects is shortened here).

How much does it cost?

$2.00 per 1,000 papers saved ($0.002 per paper), with no start fee and no monthly rental. Apify platform compute is included in the price, so you pay only for results.

• Merged duplicates count once, and papers dropped by your filters aren't charged.
• Example: the default run above (85 papers) costs about $0.17; a weekly alert with 30 new papers costs about $0.06.
• Set a maximum cost per run and the Actor stops cleanly when it reaches it.

The Pricing tab always shows the current price.

How it compares

	This Actor	Google Scholar scrapers	Single-database scrapers
Data source	Official open APIs (Crossref, arXiv, PubMed, OpenAlex)	Scholar result pages	One API
Proxies / CAPTCHAs	None	Usually needed	Usually none
Databases per run	Up to 4, merged and de-duplicated	1	1
DOI / PMID / arXiv IDs	✅ structured fields	Partial	Source-specific
Google Scholar-only results (e.g. citations of non-DOI works, author profiles)	❌ Not covered	✅	❌

Choose this Actor for reproducible, terms-friendly literature data with real identifiers. Choose a Google Scholar scraper if you specifically need Scholar's ranking or author profiles.

Speed and fair use

The Actor follows each provider's published limits:

Source	Limits followed
arXiv	1 request every 3 seconds, one connection, as its API terms require. About 100 results per request, so large arXiv pulls take a little longer.
Crossref	1 list request per second in the public pool; 3 per second if you enter your email for the polite pool.
PubMed (NCBI E-utilities)	3 requests per second, or 10 with your free NCBI API key. Every request is identified with tool=scholarly-paper-search.
OpenAlex	Without a key, all callers on the same IP share one small daily budget, so it may be skipped. Add your free key ($1/day of free usage at openalex.org) for reliable OpenAlex results and deep paging.

Data sources, terms and attribution

• Crossref: metadata from the Crossref REST API (api.crossref.org). Abstracts are supplied by publishers and may be copyrighted.
• arXiv: "Thank you to arXiv for use of its open access interoperability." Descriptive metadata is CC0. This Actor links to arXiv for PDFs and never re-hosts them. It is not endorsed by arXiv.
• PubMed: data courtesy of the U.S. National Library of Medicine, via NCBI E-utilities. NLM/NCBI disclaimer and copyright notice: https://www.ncbi.nlm.nih.gov/About/disclaimer.html. Abstracts in PubMed may incorporate material protected by copyright, so if you reproduce or redistribute them, follow the copyright holder's terms.
• OpenAlex: data is CC0 (openalex.org).
• Abstracts: to collect metadata only, turn off Include abstracts.

FAQ

Is this a Google Scholar scraper? No. It uses only open scholarly APIs. That makes it reliable and terms-friendly, and it returns structured identifiers (DOI, PMID, arXiv ID) that Google Scholar result pages don't.

Why are some citation counts empty? Citation counts come from Crossref and OpenAlex. arXiv and PubMed don't provide them, so include Crossref or add OpenAlex to get them.

Do I need API keys? No. Crossref, arXiv and PubMed work without keys. Optional keys only raise limits: an OpenAlex key gives reliable OpenAlex results, and an NCBI key makes PubMed faster.

Can I get full-text PDFs? You get pdfUrl links where the source provides them (all arXiv papers, plus open-access versions from Crossref and OpenAlex). The Actor doesn't download or re-host PDFs.

Can I export to Zotero, Mendeley or EndNote? Export CSV or JSON, then import by DOI; most reference managers can look up full metadata from a DOI list.

Can I use it with an AI agent or in a RAG pipeline? Yes. Call it through the Apify API or the Apify MCP server and feed titles, abstracts and pdfUrl links into your pipeline.