Crossref Works: Scholarly Article Metadata, DOIs and Retractions API

Crossref Works exports metadata of scholarly works (journal articles, preprints, book chapters, conference papers, datasets) registered with Crossref, the DOI agency behind 170M+ research outputs, as clean JSON, CSV or Excel. It reads the official, keyless Crossref REST API at its published rate limits. No login, no API key, no browser, no proxy.

• ✅ Search by text, author, publisher, or filter by journal ISSN, Crossref member (publisher), work type, funder and date (published, created, deposited, indexed).
• ✅ Rich records: DOI, title, authors with ORCID and affiliations (ROR), journal and ISSNs, publisher, publication dates, volume/issue/pages, citation count, references count, funders and award numbers, licences and full-text links.
• ✅ New since last run: every work registered on each full day since your previous run, without overlaps. Schedule it daily for a new-publications feed.
• ✅ Retraction feed: retraction notices with the DOIs they retract; any work carries isRetracted when a retraction notice updates it.
• ✅ No abstracts: abstracts are the publishers' copyright, so they are never requested.

Start in one minute

Click Start with the prefilled input:

{
  "mode": "search",
  "query": "large language models",
  "types": ["journal-article"],
  "fromDate": "2026-10-01",
  "untilDate": "2026-10-31",
  "maxItems": 20
}

You get: the 20 most relevant journal articles about large language models published in October 2026, with doi, title, authors, containerTitle, publisher, publishedDate, citedByCount and licenses.

Cost: $0.02: 20 works × $0.001, plus the $0.00005 actor-start fee.

Use with an AI assistant

Connect the Apify MCP server to Claude, ChatGPT, Cursor or any MCP client (https://mcp.apify.com?tools=cprussin/crossref-works loads only this actor), then paste:

Use the Apify actor cprussin/crossref-works to list papers retracted in the last week, with the retraction notice, the retracted DOI, journal and publisher.

Recipes

Daily retraction watch (scheduled task)

The first run covers yesterday; every later run covers the full days since the previous run.

{
  "mode": "retractions",
  "maxItems": 5000
}

New articles in your journals every day

{
  "mode": "newSince",
  "issns": ["1476-4687", "0036-8075"],
  "maxItems": 1000
}

Everything a publisher published last month

{
  "mode": "search",
  "memberIds": ["1968"],
  "types": ["journal-article"],
  "fromDate": "2026-10-01",
  "untilDate": "2026-10-31",
  "maxItems": 10000
}

Most-cited NSF-funded papers of 2025

{
  "mode": "search",
  "funderIds": ["100000001"],
  "fromDate": "2025",
  "untilDate": "2025",
  "sort": "is-referenced-by-count",
  "maxItems": 100
}

What can you do with Crossref data?

• Research offices and librarians: track every new paper in your journals or from your funder.
• Research-integrity teams and journalists: monitor retractions daily and check reference lists against retracted DOIs.
• Publishers and analysts: benchmark output, citations, licences and open-access share by publisher or journal.
• AI and search builders: feed clean bibliographic metadata into RAG, recommender or citation tools.

Input

Field	Description	Example
mode	search, newSince or retractions	"retractions"
query / queryAuthor / publisherName	Free text, author name, publisher name	"crispr"
issns	Journal ISSNs	["1476-4687"]
memberIds	Crossref member IDs (publishers)	["78"]
types	Work types	["journal-article"]
funderIds	Open Funder Registry IDs	["100000001"]
fromDate / untilDate	Date range (YYYY-MM-DD, YYYY-MM or YYYY)	"2026-10-01"
dateField	published, created, deposited or indexed	"created"
sort / order	e.g. is-referenced-by-count, published; desc or asc	"published"
maxItems	Max works (default 100)	1000
contactEmail	Optional: your email, for Crossref's faster polite pool	"you@example.org"

What data do you get for each work?

{
  "doi": "10.1155/ijce/9891504",
  "url": "https://doi.org/10.1155/ijce/9891504",
  "title": "RETRACTION: Analysis of Bagasse Cellulose‐Based Hydrogel for Methylene Blue Removal from Textile Industry Wastewater",
  "subtitle": null,
  "type": "journal-article",
  "authors": [
    {
      "given": "International Journal of",
      "family": "Chemical Engineering",
      "name": null,
      "orcid": null,
      "sequence": "first",
      "affiliations": []
    }
  ],
  "authorNames": ["International Journal of Chemical Engineering"],
  "containerTitle": "International Journal of Chemical Engineering",
  "shortContainerTitle": "International Journal of Chemical Engineering",
  "issns": ["1687-806X", "1687-8078"],
  "issnPrint": "1687-806X",
  "issnElectronic": "1687-8078",
  "isbns": [],
  "publisher": "Wiley",
  "memberId": "311",
  "prefix": "10.1155",
  "publishedDate": "2026-01",
  "publishedPrintDate": "2026-01",
  "publishedOnlineDate": "2026-11-01",
  "createdAt": "2026-11-01T09:51:42Z",
  "depositedAt": "2026-11-01T23:46:57Z",
  "indexedAt": "2026-11-02T00:26:28Z",
  "volume": "2026",
  "issue": "1",
  "page": null,
  "articleNumber": "9891504",
  "subjects": [],
  "referencesCount": 1,
  "citedByCount": 0,
  "funders": [],
  "licenses": [
    {
      "url": "https://creativecommons.org/licenses/by/4.0/",
      "start": "2026-11-01",
      "contentVersion": "vor",
      "delayInDays": 304
    },
    {
      "url": "http://doi.wiley.com/10.1002/tdm_license_1.1",
      "start": "2026-01-01",
      "contentVersion": "tdm",
      "delayInDays": 0
    }
  ],
  "fullTextLinks": [
    {
      "url": "https://onlinelibrary.wiley.com/doi/pdf/10.1155/ijce/9891504",
      "contentType": "application/pdf",
      "contentVersion": "vor",
      "intendedApplication": "text-mining"
    },
    {
      "url": "https://onlinelibrary.wiley.com/doi/full-xml/10.1155/ijce/9891504",
      "contentType": "application/xml",
      "contentVersion": "vor",
      "intendedApplication": "text-mining"
    },
    {
      "url": "https://onlinelibrary.wiley.com/doi/pdf/10.1155/ijce/9891504",
      "contentType": "unspecified",
      "contentVersion": "vor",
      "intendedApplication": "similarity-checking"
    }
  ],
  "updateTo": [
    {
      "doi": "10.1155/2023/2313874",
      "type": "retraction",
      "label": "Retraction",
      "source": "publisher",
      "date": "2026-11-01"
    }
  ],
  "updatedBy": [],
  "retractedDois": ["10.1155/2023/2313874"],
  "isRetracted": false,
  "updatePolicy": "https://doi.org/10.1002/crossmark_policy",
  "scrapedAt": "2026-11-02T12:00:00.000Z"
}

Titles are plain text (JATS/HTML markup removed). Abstracts and reference lists are not exported.

How much does it cost to get Crossref metadata?

Pay per event:

• $0.001 per work ($1 per 1,000).
• $0.00005 per run start. Platform compute is included.
• A tracked run with no new full day since the last run outputs nothing and costs only the start fee.

FAQ

Is this legal? Yes. Crossref states that almost all of its metadata is reusable without restriction, except abstracts, which this actor never requests; bibliographic facts are not subject to copyright and Crossref-generated data is CC0. See Crossref's metadata licensing. The actor uses the public API, identifies itself, sends one request at a time and follows the rate limit Crossref advertises (never more than 3 requests per second).

Do I need a Crossref account or key? No. Optionally give contactEmail: Crossref then serves you from its faster "polite" pool. Without it the public pool allows about one request per second; each request returns up to 1,000 works.

How does "since last run" work? In newSince and retractions modes without fromDate, each run covers the full UTC days after the previous run's last day, up to yesterday. The last day is stored in a named key-value store in your Apify account (crossref-works-state), one per filter set. If a run stops early (max items or max charge), the next run starts from the same day again.

Why retraction notices and not retracted papers? Crossref records retractions as notices that point to the retracted work (retractedDois). Any work you fetch also shows isRetracted when such a notice exists.

How many works are there per day? Typically 20,000-40,000 new DOIs a day across all types, and tens to a few hundred retraction notices.

Related actors

• Document to Markdown: turn PDFs and other documents (e.g. open-access full texts) into clean Markdown.
• Substack Scraper: newsletter posts and metadata from Substack.