📖 Crossref Scraper: Scholarly DOIs, Citations & Metadata API

Overview

Crossref Scraper pulls scholarly works from Crossref, the world's largest DOI registration agency and metadata source. Search across 150M+ journal articles, book chapters, proceedings papers, datasets, and reports by keyword, publication year, and work type. Get back DOI, title, authors, journal, publisher, year, citation count, and reference list for every matching work. No API key, and adding an email into the polite pool gets you faster and more reliable service.

It's the fastest way to build a bibliometrics dataset, feed a RAG system with authoritative scholarly metadata, or run a literature-review pipeline.

Reliability posture: blocked, empty, or failed runs are never charged. You only pay for a work record that was actually delivered.

✅ No API key required | ✅ 150M+ works | ✅ Polite-pool option | ✅ MCP-ready for AI agents

Features

150M+ works. Every DOI registered with Crossref across every publisher. Year range. Filter by from-year and to-year for a targeted window. Work-type filter. journal-article, book-chapter, proceedings-article, dataset, report, and more. Citation counts. Reference count and cited-by count on every record. Polite pool. Add a contact email for faster and more reliable service from Crossref.

How it works

The actor calls Crossref's public works endpoint with your query, year filters, and work-type filter. Results are paginated using Crossref's deep-pagination cursor so runs can scale into the tens of thousands of records without hitting an offset limit.

Every work is normalised to a flat JSON row with DOI, title, authors, journal, year, publisher, and citation metrics. If you supply a contact email in the mailto field the actor uses Crossref's "polite pool" endpoint, which is significantly faster and more reliable than the anonymous public endpoint.

🧾 Input configuration

{
  "query": "large language models",
  "fromYear": 2023,
  "toYear": 2026,
  "workType": "journal-article",
  "maxResults": 500,
  "mailto": "you@example.com"
}

📤 Output format

Captured from a live run against Crossref (query large language models, work type journal-article, from year 2023):

{
  "doi": "10.1162/coli_a_00558",
  "title": "Large Language Models Are Biased Because They Are Large Language Models",
  "type": "journal-article",
  "year": 2025,
  "published_date": "2025",
  "authors": ["Philip Resnik"],
  "journal": "Computational Linguistics",
  "publisher": "MIT Press",
  "cited_by_count": 24,
  "references_count": 92,
  "issn": ["0891-2017", "1530-9312"],
  "url": "https://doi.org/10.1162/coli_a_00558",
  "scraped_at": "2026-07-15T04:14:41.192Z"
}

Every work record contains these fields:

Field	Description
🆔 doi	Digital Object Identifier
📄 title	Work title
🏷️ type	Publication type (journal-article, book-chapter, etc.)
📅 year	Publication year
📅 published_date	Publication date, precision depends on what Crossref has on file (year, or year and month, or full date)
👥 authors	Array of author names
📰 journal	Journal or container title
🏛️ publisher	Publisher name
🔗 cited_by_count	Number of works that cite this one
📈 references_count	Number of references in the work
🔢 issn	Array of journal ISSNs
🌐 url	Canonical DOI URL
🕒 scraped_at	ISO timestamp of when the record was captured

💼 Common use cases

Bibliometrics & research metrics Compute citation distributions by year, journal, or topic across a corpus. Rank the most-cited works in a niche for a report or grant proposal.

Literature reviews & RAG Feed authoritative scholarly metadata into a retrieval-augmented generation system. Build a searchable in-house index of publications across a research area.

Institutional research support Pull an author's publication list for a CV or grant application. Monitor a lab's output year over year.

Competitive research intel Track which journals a competitor publishes in and how often. Watch new works in a niche on a scheduled cadence.

🚀 Getting started

1. Open the actor and enter a query (e.g. crispr, large language models).
2. Optionally set fromYear and toYear for a targeted window.
3. Optionally filter by workType (journal-article, book-chapter, dataset).
4. Set maxResults to control cost.
5. Add a contact email in mailto for the polite-pool speed boost.
6. Click Start. Download as JSON, CSV, or Excel, or pull via API or MCP.

FAQ

What is the "polite pool"? Crossref serves anonymous requests best-effort and reserves faster infrastructure for callers that identify themselves with a contact email. Adding your email in mailto opts into that pool and typically yields lower latency and higher reliability.

How does deep pagination work? Crossref supports a cursor-based deep pagination that avoids the offset cap on the standard rows/offset API. This actor uses cursor pagination automatically so runs can scale into tens of thousands of records for a single query.

How much does it cost? Pay per work returned, pay as you go. No subscription, no monthly minimum.

Can I use it in an AI agent? Yes. It's exposed as an MCP tool. See below.

Use in Claude, ChatGPT & any MCP agent

https://mcp.apify.com/?tools=themineworks/crossref-scholarly-metadata

Or call it programmatically with the Apify client:

import { ApifyClient } from 'apify-client';

const client = new ApifyClient({ token: 'YOUR_APIFY_TOKEN' });

const run = await client.actor('themineworks/crossref-scholarly-metadata').call({
  query: 'large language models',
  fromYear: 2023,
  workType: 'journal-article',
  maxResults: 100,
});

const { items } = await client.dataset(run.defaultDatasetId).listItems();
console.log(items);

🛠️ Complete your research intel pipeline

Got the metadata. Now widen the corpus:

• arXiv Paper Scraper: pull preprints on the same topic before they hit a journal.
• PubMed NCBI Scraper: pull biomedical literature with MeSH terms.
• OpenCitations: pull the full citation graph for any DOI.

Typical flow: crossref finds the works, arxiv adds preprints, opencitations builds the citation graph.

Questions or need a custom field set? Reach out through the Apify profile.