Crossref Works Scraper

Try a live sample

Open the research-paper metadata Task to inspect a small editable query and DOI-backed output before scaling.

Search Crossref works or enrich a DOI list, then export publication metadata to CSV or JSON for literature reviews, citation workflows, and library enrichment. This Actor returns research-paper metadata, not paper PDFs or full text.

Use this actor when you need clean DOI-backed publication rows without maintaining your own Crossref integration. It is useful for publishers, universities, academic databases, AI/RAG teams, citation tools, research analysts, and market-intelligence workflows that need publication metadata at predictable cost.

What it returns

• DOI, title, abstract, publisher, journal/container title
• Authors, publication dates, type, subject
• Reference count, citation count, indexed date, license records
• DOI URL and the query or DOI that produced the row

Common use cases

• Enrich DOI lists with titles, authors, journal names, publishers, and dates
• Build scholarly datasets for RAG, literature reviews, or semantic search
• Collect results for publisher, journal, author, or keyword research
• Power citation workflows, library metadata cleanup, and research market maps
• Feed publication metadata into AI agents for summarization and deduplication

Your first run

An Actor is the data-collection tool; a run is one execution. Its Dataset contains the returned rows. A saved Task keeps an input configuration for reuse.

1. Open the sample Task above or this Actor's Input tab.
2. Start with the ten-result query below. Review the live Pricing tab and your run's charge limit before starting.
3. When the run finishes, inspect its output Dataset. Check titles and DOIs for relevance before downloading CSV or JSON.
4. Save a Task once the input suits your workflow. Increase the query list only after checking the first output and charge.

Example input: keyword search

Provide search queries, DOIs, or both. Crossref works best with publication titles, author names, journal names, or exact DOI values.

{
  "queries": ["retrieval augmented generation"],
  "dois": [],
  "maxResults": 10
}

Example input: one DOI

For DOI enrichment only, explicitly leave queries empty so the form's default search does not add unrelated results. Supply a DOI identifier, not a publisher-page URL.

{
  "queries": [],
  "dois": ["10.1038/s41586-020-2649-2"],
  "maxResults": 1
}

Input limits

Field	Meaning
queries	Up to 25 search strings, each up to 500 characters.
dois	Up to 100 DOI strings, each up to 300 characters.
maxResults	Results requested per query, from 1 to 100; default 50. It does not limit the number of DOI lookups.

maxResults is not a global result or spending cap. Two queries with maxResults: 10 plus five DOIs can return up to 25 rows before deduplication. Use a small input and Apify's run charge limit to bound your first attempt. If both input lists are empty, the Actor runs a small default search for retrieval augmented generation rather than doing nothing.

The Actor requests one result page per query. It does not exhaustively paginate Crossref or restrict results to papers published since your previous run. Repeated runs can return and charge for the same works again; deduplication applies within a run, not across runs.

Output fields

Every returned row is saved to the default Apify Dataset with these field names:

Fields	Content
doi, url	DOI identifier and publication/DOI URL.
title, subtitle, abstract	Title, subtitle, and abstract when supplied by Crossref. Abstracts may contain markup.
publisher, type, containerTitle	Publisher, work type, and journal or other container title.
authors	List of author objects with name, orcid, and sequence.
issn, isbn, subjects	Identifier and subject lists.
publishedAt, createdAt, indexedAt	Formatted dates when present. Missing month or day components are filled with 01; do not treat that as verified day-level precision.
referenceCount, isReferencedByCount	Crossref's reference and citation counts when present.
licenses	Crossref license records, not a list of URL strings.
sourceType, source	Whether the row came from a query or DOI lookup, and the original input string.

Missing scalar metadata can be null; missing lists are empty. A missing abstract does not mean the paper is missing. Crossref coverage depends on deposited metadata, and a citation count here is not a Google Scholar citation count.

The Actor retries temporary Crossref rate limits and upstream errors. Invalid or missing DOIs are reported in the non-billable RUN_SUMMARY key-value-store record instead of being written as paid dataset rows. One bad query or DOI does not cancel successful inputs in the same run.

SEO keywords

Crossref scraper, DOI metadata scraper, Crossref works API, scholarly metadata export, academic publication scraper, citation data scraper, DOI to CSV, research paper metadata API.

Pricing

Check the live Pricing tab for the current per-result rate before running. Charges use the configured dataset-item event, not the number of search terms. This guide does not change the live price. Failed-input diagnostics are recorded separately in RUN_SUMMARY, not as dataset rows.

Troubleshooting

• Fewer results than requested: Crossref may have fewer matches, duplicate works are removed within the run, or individual inputs may have failed. maxResults is a requested maximum, not a guaranteed count.
• A DOI was skipped: Check the run's key-value store record named RUN_SUMMARY. It includes rowsSaved, queriesRequested, doisRequested, and failedInputs with the affected input and error. Confirm the DOI exists in Crossref; not all DOI registration agencies use Crossref.
• The run succeeded but the output is empty: Inspect RUN_SUMMARY and the log. A completed process alone does not guarantee useful rows. Correct the input or wait for an upstream outage to clear before another attempt.
• I need PDFs or every matching paper: This Actor provides a bounded metadata lookup. It does not download papers or provide exhaustive search pagination.

Responsible use

This actor extracts public metadata from Crossref. Respect Crossref API etiquette, publisher rights, and license terms for downstream use. Metadata can be incomplete or updated; verify critical bibliographic records before using them in formal publication, compliance, or procurement workflows.

Automation ideas

Use saved Tasks, schedules, or the Apify API to send DOI batches into a citation manager, spreadsheet, or research pipeline. AI agents can summarize available abstracts and cluster topics while preserving DOI source links; do not invent missing authors, abstracts, or dates. In an Apify MCP workflow, select this Actor, pass the same input fields shown above, and inspect the Dataset before using it as evidence. Configure your AI client's connection separately; this Actor does not set it up for you.