OpenAlex Scraper

What does OpenAlex Scholarly Data Scraper do?

OpenAlex Scholarly Data Scraper collects research-paper metadata, author profiles, institutional records, and publication-source details from OpenAlex. Choose an entity, enter a search term, and set a result limit to create a dataset for literature discovery, bibliometric analysis, or research monitoring.

For research works, the output includes titles, author names, affiliations, publication years, DOIs, citation counts, and abstracts when available. Download the results as JSON, CSV, Excel, or XML, or connect them to an Apify integration.

Why use OpenAlex Scholarly Data Scraper?

• Literature discovery - Collect paper titles, publication dates, DOIs, and available abstracts for an initial review of a research topic.
• Citation analysis - Sort results by citation count to identify frequently cited works or profiles within your search results.
• Researcher and institution lookup - Find named authors and organizations without manually copying profile details.
• Repeatable collection - Reuse the same inputs for scheduled runs and compare downloaded datasets over time.
• Batch output - Each completed page is saved during the run, so earlier results remain available if a later request fails.
• Bounded recovery - Temporary failures receive a limited number of retries rather than keeping the run waiting indefinitely.

The actor does not guarantee uninterrupted access, complete abstracts, or a particular number of matching records. Results depend on OpenAlex coverage and the available anonymous-access budget.

What data can you extract from OpenAlex?

Entity	Data collected	Search guidance
works	Paper titles, author names, affiliations, publication years, DOIs, abstracts, concepts, and citation counts	Use a research term or paper title
authors	Display names, publication counts, citation counts, one last-known institution, and ORCID	Use an author's name; this is not a subject-expertise filter
institutions	Organization names, countries, institution types, publication counts, and citation counts	Use an institution name
sources	Journal or other source names, publisher/host organization names, ISSNs, counts, and open-access indicators	Use a publication-source name
concepts	Concept names, levels, descriptions, publication counts, and citation counts	Legacy subject classification; OpenAlex has deprecated concepts

The input also retains venues as a legacy option. Use sources for new publication-source searches; the actor sends the selected entity as supplied and does not translate legacy endpoint names. Topics, standalone publisher records, full-text downloads, and advanced geographic or date filters are not exposed by this actor.

How to use the actor

1. Open the actor's input form in Apify Console.
2. Select an entity, such as works for research papers.
3. Enter a search term appropriate to that entity.
4. Set results_wanted and, for larger collections, max_pages.
5. Choose relevance or citation sorting, then start the run.
6. Review the dataset preview and download your preferred format.

Start with 20 results to check whether your query produces useful records. The Console shows machine learning as a search example. Omitted or blank search input lists records without a keyword filter; it does not automatically run that example search.

Input Parameters

Parameter	Type	Required	Default	Description
search	String	No	Empty	Research term, paper title, author name, institution name, or source name, depending on the selected entity
entity	String	No	"works"	One of works, authors, institutions, sources, concepts, or the legacy venues option
results_wanted	Integer	No	20	Maximum number of records to save; minimum 1
max_pages	Integer	No	10	Maximum number of result pages to request; minimum 1
sort	String	No	"relevance_score:desc"	Sort expression; use cited_by_count:desc for descending citation count. Relevance sorting is omitted when search is blank.

A page contains at most 100 records. Both result and page limits apply: results_wanted: 500 with max_pages: 2 can collect at most 200 records, and fewer when matches or access are limited. OpenAlex's basic paging limit is 10,000 results; this actor does not provide cursor continuation beyond that limit.

There is no API-key or proxy input in the current version. Adding unrecognized credential or proxy properties to the input does not enable those features.

Output Data

Fields depend on the selected entity. Missing scalar metadata may be null; missing names or classifications may be empty arrays. Citation and publication counts describe the source data at collection time and can change later.

Field	Type	Applies to	Description
id	String or null	All	OpenAlex entity identifier
url	String or null	All	Entity URL derived from its identifier
source	String	All	Always openalex.org
title	String or null	Works	Work title
authors	Array	Works	Author display names
institutions	Array	Works	Institution names from authorships; duplicates can occur
publication_year	Integer or null	Works	Publication year
doi	String or null	Works	DOI URL when available
abstract	String or null	Works	Available abstract reconstructed as readable text; not the full paper
concepts	Array	Works	Legacy concept display names
cited_by_count	Integer	All	Citation count
type	String or null	Works, institutions, sources/venues	Record type as supplied by OpenAlex
display_name	String or null	Non-work entities	Entity name
works_count	Integer	Non-work entities	Associated work count
last_known_institution	String or null	Authors	First last-known institution when several are present
orcid	String or null	Authors	Researcher's ORCID identifier
country_code	String or null	Institutions	Country code
publisher	String or null	Sources/venues	Host organization name, or publisher value when available
issn_l	String or null	Sources/venues	Linking ISSN
issn	Array or null	Sources/venues	ISSNs, with linking ISSN as a fallback when available
is_oa	Boolean	Sources/venues	Open-access indicator
is_in_doaj	Boolean	Sources/venues	Directory of Open Access Journals indicator
level	Integer or null	Concepts	Legacy hierarchy level; the current mapping returns null for a zero level
description	String or null	Concepts	Available concept description

Usage Examples

Search research papers

Collect up to 20 works related to machine learning, ordered by relevance.

{
  "search": "machine learning",
  "entity": "works",
  "results_wanted": 20
}

Look up an author

Search author names and rank matching profiles by citation count.

{
  "search": "Yoshua Bengio",
  "entity": "authors",
  "sort": "cited_by_count:desc",
  "results_wanted": 5
}

Find an institution

Collect matching institutional records for Stanford University.

{
  "search": "Stanford University",
  "entity": "institutions",
  "results_wanted": 5
}

Collect a larger paper dataset

Request up to 300 climate-change works across at most three pages, ranked by citation count.

{
  "search": "climate change",
  "entity": "works",
  "sort": "cited_by_count:desc",
  "results_wanted": 300,
  "max_pages": 3
}

Sample Output

This work example uses metadata checked on September 30, 2026. Citation counts and affiliations may change. The abstract was unavailable for this record, so its value is null.

{
  "id": "https://openalex.org/W2919115771",
  "url": "https://openalex.org/W2919115771",
  "source": "openalex.org",
  "title": "Deep learning",
  "authors": [
    "Yann LeCun",
    "Yoshua Bengio",
    "Geoffrey E. Hinton"
  ],
  "institutions": [
    "Meta (United States)",
    "New York University",
    "Université de Montréal",
    "Google (United States)",
    "University of Toronto"
  ],
  "publication_year": 2015,
  "doi": "https://doi.org/10.1038/nature14539",
  "abstract": null,
  "concepts": [
    "Computer science",
    "Deep learning",
    "Artificial intelligence",
    "Abstraction",
    "Representation (politics)",
    "Layer (electronics)",
    "Object (grammar)",
    "Backpropagation",
    "Convolutional neural network",
    "Feature learning",
    "Pattern recognition (psychology)",
    "Speech recognition",
    "Artificial neural network",
    "Epistemology",
    "Politics",
    "Philosophy",
    "Political science",
    "Organic chemistry",
    "Chemistry",
    "Law"
  ],
  "cited_by_count": 85092,
  "type": "article"
}

Tips for Best Results

• Match the query to the entity - Use topics for works, people’s names for authors, and organization names for institutions.
• Check small runs first - Inspect 20 results before requesting hundreds of records.
• Choose the ranking deliberately - Citation sorting emphasizes frequently cited records, not necessarily the newest or most relevant research.
• Allow enough pages - Each page supplies at most 100 records; a low page limit can stop collection before the result target.
• Plan for missing data - An absent abstract, DOI, ORCID, or affiliation is not necessarily an extraction failure.
• Avoid overlapping large runs - Repeated anonymous requests can exhaust the available budget. Delay subsequent runs when OpenAlex reports a limit.
• Keep your earlier results - A failed later page does not remove batches already saved to the dataset.

Integrations

Use Apify's integrations and dataset access to connect the results to your workflow:

• Google Sheets or Airtable - Review citations, authors, and institutions in a spreadsheet or database.
• Make or Zapier - Connect completed runs to downstream research workflows.
• Slack - Send notifications through a configured integration.
• Webhooks - Trigger processing when a run completes or fails.
• Programmatic dataset access - Retrieve results for notebooks, dashboards, or a research database.

Export Formats

Download datasets as JSON, CSV, Excel, or XML. JSON preserves arrays such as author and institution lists for applications that need the original record structure.

Frequently Asked Questions

Is an API key required?

No. The current actor uses keyless access and does not accept an API key. Anonymous access has a limited daily budget, so a run can be throttled even when its input is valid.

What happens when OpenAlex returns a rate-limit error?

The actor makes at most four attempts per page: the initial request and three retries. It follows valid server-provided retry timing or uses short increasing waits with a small random variation when no timing is provided.

The run fails when retries are exhausted, a rate-limit response explicitly reports no remaining daily budget, or the requested cooldown exceeds 60 seconds. Malformed responses and permanent access or input errors are not retried. Wait for the limit to reset before starting another run; completed batches remain in the dataset.

Will a residential proxy solve HTTP 429 errors?

Not necessarily. A residential proxy may help when throttling is specific to the current network address, but it is not a guaranteed fix for an exhausted upstream budget or a service-wide restriction. It also adds cost and may increase response time.

This actor currently has no proxy integration. A proxy has not been tested against the reported cloud failure, so it is not presented as a verified solution. Respect OpenAlex's usage limits rather than relying on address rotation to avoid them.

Can I collect journals and publication venues?

Yes. Use entity: "sources" to search publication-source records. The publisher output describes the available host organization or publisher name; it is not a separate publisher profile. Prefer sources over the legacy venues input.

Are abstracts the full text of papers?

No. The actor returns an abstract only when OpenAlex supplies one. It does not download the article's full text or create a summary when an abstract is missing.

Can I find researchers by their expertise?

Not directly. Author searches match names. Searching authors for a subject such as artificial intelligence does not provide an expertise filter. Search works by subject and inspect their author names for an initial literature-based workflow.

How many records can I collect?

Collection stops at your result target, page limit, end of matching results, or an unrecoverable error. Each page contains at most 100 records, and basic paging cannot go beyond OpenAlex's 10,000-result limit. A requested count is a maximum, not a guarantee.

Can I schedule recurring runs?

Yes. Configure an Apify schedule with the desired inputs. Allow enough time between runs to avoid repeatedly consuming the anonymous budget, and monitor failures before scheduling larger collections.

What should I do if fields are empty?

Check several records and the entity type. Some fields apply only to works or profiles, and OpenAlex may not publish a value for a particular record. Missing metadata is represented by null values, empty arrays, or the mapped count/boolean defaults.

Support

Use the actor's Issues tab or contact the developer through Apify Console. Include the run URL, selected entity, result and page limits, and the relevant error message. Do not share private credentials.

Resources

• Apify dataset documentation
• Apify schedules
• OpenAlex Help Center

Legal Notice

This actor collects publicly available scholarly metadata from OpenAlex. You are responsible for complying with OpenAlex's terms, applicable laws, and any requirements of your downstream workflow. Respect usage limits and handle author or institutional information responsibly. Metadata availability does not imply permission to download or redistribute the full text of a publication.