Scholarly Papers API: OpenAlex & Crossref Search, DOI Lookup

Search scholarly papers, preprints and books through OpenAlex and Crossref instead of scraping Google Scholar, for researchers, librarians, literature-review tools and AI agents. It costs $0.002 per work, so 1,000 works cost $2. Results from both sources are merged by DOI and billed once.

Search the scholarly record without scraping Google Scholar. OpenAlex (250M+ works, open data, CC0) and Crossref (publisher-registered DOI metadata) answer keyword queries with filters for year, work type and open access, sorted by relevance, citations or date. Every work comes back as one row: title, authors with affiliations, year, venue, publisher, DOI, open-access status and PDF link, citation counts and field-weighted impact, topic, keywords and (optionally) the abstract. Look up lists of DOIs in bulk. Results from both sources are merged by DOI and billed once.

Built and maintained by Lathewright. Lathewright is operated by AI systems under human ownership. A person reviews and approves everything we publish, send and sell. Pay only for what you get: $0.002 per work.

What you can do with it

• Literature monitoring: newest papers on a topic every week, with open-access links, into a reading list or a summarizer.
• Research agents: give an LLM citation-backed sources with DOIs instead of hallucinated references.
• Citation and impact analysis: most-cited works, field-weighted citation impact, retraction flags, author affiliations and countries.
• Reference enrichment: turn a list of DOIs into full metadata and PDF links.

Example tasks

Paste any of these into the Input tab, or send the same JSON through the API. Prices are the per-result prices listed under Pricing.

1. Most cited open-access papers on retrieval-augmented generation

{
  "mode": "search",
  "query": "retrieval-augmented generation",
  "sources": ["openalex"],
  "fromYear": 2024,
  "sortBy": "citations",
  "openAccessOnly": true,
  "maxResultsPerSource": 20
}

What comes back: One row per work with title, authors, year, venue, citedByCount, isOpenAccess, doi, pdfUrl and url. Up to 20 rows, at most $0.04.

2. Look up a DOI

{
  "mode": "doiLookup",
  "dois": ["10.1038/s41586-024-07487-w"]
}

What comes back: The work's metadata row, or a failed-lookup row if the DOI is unknown.

3. Preprints with abstracts from both sources

{
  "mode": "search",
  "query": "large language models evaluation",
  "sources": ["openalex", "crossref"],
  "workTypes": ["preprint"],
  "includeAbstract": true,
  "maxResultsPerSource": 15
}

What comes back: Preprints merged by DOI, each with the abstract field filled when the source has one.

Input

Field	Type	Notes
mode	select	search or doiLookup
query	string	Keywords; quotes for phrases
sources	array	openalex, crossref (default both)
fromYear, toYear	integer	Publication year window
workTypes	array	article, preprint, review, chapter, book, conference, dataset, dissertation, report
openAccessOnly	boolean	Free full text (OpenAlex) or open licence (Crossref)
sortBy	select	relevance, citations, date
includeAbstract	boolean	Abstract text when available
maxResultsPerSource	integer	Cap per source (default 50)
dois	array	For doiLookup
openalexApiKey	secret	Optional free key; without it requests share a small per-IP daily budget
contactEmail	string	Optional; sent as mailto to join both polite pools

{ "mode": "search", "query": "large language models evaluation", "fromYear": 2024, "workTypes": ["article", "preprint"], "sortBy": "citations", "includeAbstract": true, "maxResultsPerSource": 100, "openalexApiKey": "<optional>" }

{ "mode": "doiLookup", "dois": ["10.1038/s41586-021-03819-2", "10.48550/arXiv.2303.08774"], "includeAbstract": true }

Output

{
  "type": "work",
  "source": "openalex",
  "id": "https://openalex.org/W4399054296",
  "doi": "10.1038/s41586-024-07487-w",
  "title": "Accurate structure prediction of biomolecular interactions with AlphaFold 3",
  "authors": [{ "name": "Josh Abramson", "position": "first", "institutions": ["Google DeepMind"], "countries": ["GB"] }],
  "authorCount": 48,
  "year": 2024,
  "publicationDate": "2024-05-08",
  "workType": "article",
  "language": "en",
  "venue": "Nature",
  "issn": "0028-0836",
  "publisher": "Nature Portfolio",
  "volume": "630", "issue": "8016", "pages": "493-500",
  "isOpenAccess": true,
  "oaStatus": "hybrid",
  "oaUrl": "https://www.nature.com/articles/s41586-024-07487-w.pdf",
  "landingPageUrl": "https://doi.org/10.1038/s41586-024-07487-w",
  "pdfUrl": "https://www.nature.com/articles/s41586-024-07487-w.pdf",
  "license": "cc-by",
  "citedByCount": 15516,
  "referencedWorksCount": 65,
  "fwci": 412.3,
  "citationPercentile": 99.99,
  "isRetracted": false,
  "topic": "Protein Structure and Dynamics",
  "subfield": "Structural Biology",
  "field": "Biochemistry, Genetics and Molecular Biology",
  "keywords": ["AlphaFold", "Protein structure prediction"],
  "abstract": "The introduction of AlphaFold 2 has spurred a revolution...",
  "pmid": "38718835",
  "url": "https://doi.org/10.1038/s41586-024-07487-w"
}

Crossref rows carry the same fields with source: "crossref" (no topics, impact metrics or OA URLs, but publisher-registered licences and PDF links where deposited). DOIs that neither source knows are type: "error" rows at the reduced rate.

Pricing

Event	Price	Notes
Work found (work)	$0.002	Primary event. 500 works = $1.00; a DOI seen in both sources is billed once
Failed lookup (lookup-error)	$0.0005	DOI not found

Plus Apify's standard start fee ($0.00005 per run). No proxies, no browser. Set a maximum cost per run and the Actor stops cleanly at the limit.

How to use it

• Apify Console: fill in the form, press Start, download JSON, CSV or Excel from the Output tab.
• API: POST https://api.apify.com/v2/acts/lathewright~scholarly-search/run-sync-get-dataset-items?token=<token> with the input JSON as the body.
• n8n, Make, Zapier: use the Apify node/module, pick this Actor, paste the input JSON.

Use from an AI agent (MCP)

Add this Actor as a tool in Claude, Cursor, VS Code or any MCP client with https://mcp.apify.com/?tools=lathewright/scholarly-search. The Actor description is the tool description, and the input fields above are the tool parameters. Example prompts: "Find the most cited 2025 papers on retrieval-augmented generation with open-access PDFs", "Get metadata for these 30 DOIs".

Data sources, robots.txt and fair use

Checked on 19 October 2026 for the exact paths the Actor requests:

• OpenAlex API (api.openalex.org/works): robots.txt allows the path. OpenAlex data are CC0. Since 2025 OpenAlex meters requests: without a key, all callers behind one IP address share a small free daily budget (about 100 requests), which can run out on shared infrastructure, so for anything beyond a quick query add your free openalexApiKey (openalex.org/settings/api) with its own daily budget. The Actor sends mailto and at most five requests per second (OpenAlex allows ten).
• Crossref REST API (api.crossref.org/works): no robots.txt is served; Crossref's API terms allow public use and ask clients to identify themselves (the Actor sends mailto in the query and in its User-Agent to join the polite pool) and to stay around one request per second, which the Actor does. Crossref returns HTTP 429 when a network sends too much traffic; the Actor then stops that source with a warning and keeps the OpenAlex results. During local testing from a shared network Crossref answered the first requests and rate-limited the rest, so its live behaviour on the platform should be checked after publishing.
• PubMed E-utilities and the arXiv API are deliberately not used: their robots.txt files disallow the API paths for generic crawlers. OpenAlex indexes both corpora, and PubMed ids are returned in pmid where known.
• Author names and affiliations are the published bibliographic record of each work; no contact details, ORCID profiles or emails are collected.

FAQ

Why do some abstracts come back empty? Publishers do not always deposit abstracts; OpenAlex reconstructs them from an inverted index when available and Crossref returns them only for a minority of records.

Why are counts different between sources? OpenAlex counts citations across its whole corpus (including preprints and repositories); Crossref counts only DOI-to-DOI references registered by publishers.

Can I filter by author or institution? Put the name in query for a relevance match; structured author and institution filters are a planned addition.