Research Papers: Search PubMed, arXiv & Crossref

Search scientific papers in three big databases at once and get one clean, uniform record per paper:

• PubMed: about 38 million biomedical and life-science articles, with MeSH terms
• arXiv: preprints in physics, mathematics, computer science, AI, statistics, economics and more
• Crossref: metadata of almost every journal article, book chapter and conference paper with a DOI, with citation counts

Every record has title, abstract, authors, journal, publication date, DOI, keywords or categories, links to the paper, full text and PDF where available, and an open-access flag. A paper found in several databases is returned once.

Or use lookup mode to turn a list of DOIs, PubMed IDs and arXiv IDs into full records, for example to enrich a reference list.

• ✅ One schema for all sources: no more mapping three different APIs
• ✅ Date filters and newest-first sorting, for literature monitoring with a schedule
• ✅ Deduplicated by DOI across sources
• ✅ Advanced queries: exact phrases, PubMed field tags, arXiv fields like ti:, au:, cat:cs.LG
• ✅ Official APIs used within their rate limits; no key needed
• ✅ Fair billing: per returned paper

Input examples

Newest papers on a topic since 2025, from all sources:

{
    "query": "crispr base editing",
    "from": "2025",
    "sortBy": "date",
    "maxResultsPerSource": 250
}

AI preprints and journal articles:

{
    "query": "\"large language model\" agents",
    "sources": ["arxiv", "crossref"],
    "maxResultsPerSource": 400
}

Look up known papers:

{
    "mode": "lookup",
    "ids": ["10.1038/s41586-020-2649-2", "42900733", "arXiv:1706.03762"]
}

Output examples

Real output from a test run (abstracts and author lists shortened). From Crossref:

{
    "source": "crossref",
    "doi": "10.1038/s41586-020-2649-2",
    "title": "Array programming with NumPy",
    "abstract": "Array programming provides a powerful, compact and expressive syntax for accessing, manipulating and operating on data in vectors, matrices and higher-dimension…",
    "authors": ["Charles R. Harris", "K. Jarrod Millman", "Stéfan J. van der Walt", "…"],
    "journal": "Nature",
    "publisher": "Springer Science and Business Media LLC",
    "publicationDate": "2020-09-16",
    "year": 2020,
    "type": ["journal-article"],
    "citationCount": 25397,
    "url": "https://doi.org/10.1038/s41586-020-2649-2",
    "pdfUrl": "https://www.nature.com/articles/s41586-020-2649-2.pdf",
    "license": "https://creativecommons.org/licenses/by/4.0",
    "openAccess": true
}

From PubMed:

{
    "source": "pubmed",
    "doi": "10.1073/pnas.2607793123",
    "pmid": "42900733",
    "title": "Evolution of the arbuscular mycorrhizal regulatory network through differential GRAS gene retention and protein interaction in land plants.",
    "abstract": "Plants rely on an ancient genetic \"toolkit\" to establish arbuscular mycorrhizal symbiosis (AMS) with fungal partners. …",
    "authors": ["Wen-Shen Wu", "Peng Zhao", "Zhen Zeng", "…"],
    "journal": "Proceedings of the National Academy of Sciences of the United States of America",
    "publicationDate": "2026-10-27",
    "type": ["Journal Article"],
    "keywords": ["Mycorrhizae", "Symbiosis", "Phylogeny", "Solanum lycopersicum"],
    "url": "https://pubmed.ncbi.nlm.nih.gov/42900733/",
    "fullTextUrl": "https://doi.org/10.1073/pnas.2607793123"
}

From arXiv:

{
    "source": "arxiv",
    "doi": "10.48550/arxiv.1706.03762",
    "arxivId": "1706.03762",
    "version": 7,
    "title": "Attention Is All You Need",
    "abstract": "The dominant sequence transduction models are based on complex recurrent or convolutional neural networks in an encoder-decoder configuration. …",
    "authors": ["Ashish Vaswani", "Noam Shazeer", "Niki Parmar", "…"],
    "publicationDate": "2017-06-12",
    "updated": "2023-08-02",
    "keywords": ["cs.CL", "cs.LG"],
    "primaryCategory": "cs.CL",
    "url": "https://arxiv.org/abs/1706.03762",
    "pdfUrl": "https://arxiv.org/pdf/1706.03762v7",
    "openAccess": true
}

A summary per source (results, total matches, duplicates skipped) is saved as OUTPUT in the key-value store.

Pricing

A fixed price per returned paper. In our test 498 papers from all three sources took 17 seconds with the default 256 MB memory.

Paid Apify plans pay less: 10% off on Starter, 20% on Scale and 30% on Business and higher plans.

FAQ

Why is arXiv slower? arXiv asks for at most one request every three seconds. The Actor follows this; each request returns up to 200 papers.

Why does PubMed have no citation counts? PubMed does not provide them. Crossref does (citationCount, the number of works citing the paper that are registered with Crossref).

Can I get the full text? The Actor returns metadata and abstracts plus links: pdfUrl (arXiv and many open-access journals) and fullTextUrl (PubMed Central or the publisher page). Use and share full texts according to their license.

Why do I need a contact email? You don't. NCBI and Crossref ask tools to send one; with it, Crossref serves requests faster. It is sent to these services only.

More tools from Tidelane Data

• PDF to Markdown: PDFs to Markdown and text, with tables and OCR for scans
• Hacker News Scraper: search, mentions and Who Is Hiring jobs
• RSS Feed Reader: news, blogs and podcasts to JSON
• SEC EDGAR: filings, financial statements and full-text search

All our tools: apify.com/tidelane

---

Impressum

Levi Griesinger, Tidelane Data
Besigheimer Straße 18
74369, Löchgau, Deutschland
E-Mail: LeviGriesinger@gmail.com