Data Sources

PRIMA Scholar searches ten academic databases through its search server. A search reaches all ten unless you name a smaller set, and PRIMA Scholar leaves CORE out when you have no CORE key. This page explains what each database covers and how PRIMA Scholar treats its results.

At a Glance

Database	Name in sources	Coverage	Key Needed	Built-In Limit
OpenAlex	openalex	250 million or more works, all disciplines	No	10 requests a second
PubMed	pubmed	Biomedical and life sciences	Optional	3 a second, or 10 with a key
arXiv	arxiv	Preprints in physics, maths, computer science, biology, finance and statistics	No	1 request every 3 seconds
Semantic Scholar	semantic_scholar	All disciplines, with a citation graph	Optional	100 requests every 5 minutes
CrossRef	crossref	DOI records, all disciplines	No	50 requests a second
CORE	core	300 million or more open access papers	Yes, free	10 requests a second
Europe PMC	europe_pmc	Biomedical and life sciences, with European research	No	10 requests a second
ERIC	eric	Education research	No	10 requests a second
bioRxiv/medRxiv	biorxiv	Biology and medicine preprints	No	5 requests a second
DBLP	dblp	Computer science bibliography	No	5 requests a second

The coverage figures come from the PRIMA Scholar README. The search server holds the limits in its own code and waits whenever a database reaches its limit.

How Results Are Combined

PRIMA Scholar sends your query to every chosen database at the same time and gathers the answers. When two databases return the same paper with the same DOI, it keeps the copy with the most complete details. It never merges papers without a DOI, so the same paper can appear twice when one database has no DOI for it.

A database that fails does not stop the search. Its error appears in the errors part of the response, and the other databases’ results come back as usual.

When you set a date window, PRIMA Scholar passes it to each database that can use one. It then checks every result against the window again on your computer. It keeps a paper that has only a year when that year falls inside the window, and it keeps a paper with no date at all. A year range set on its own, with no exact date window, is not checked again on your computer.

Choosing Databases by Discipline

The search wizard reads the words in your query and suggests databases for the discipline it detects. It looks for keywords linked to each field and suggests the primary databases below.

Discipline	Suggested Databases
Psychology	OpenAlex, Semantic Scholar, PubMed
Education	ERIC, OpenAlex, Semantic Scholar
Neuroscience	PubMed, Europe PMC, bioRxiv/medRxiv
Business and Management	OpenAlex, Semantic Scholar, CrossRef
Computer Science and AI	Semantic Scholar, arXiv, DBLP
Philosophy and Humanities	OpenAlex, CrossRef
Biomedical and Life Sciences	PubMed, Europe PMC, bioRxiv/medRxiv
Engineering	OpenAlex, Semantic Scholar, arXiv
Social Sciences	OpenAlex, Semantic Scholar, CrossRef
Mathematics and Physics	arXiv, Semantic Scholar, OpenAlex
Economics	OpenAlex, CrossRef, Semantic Scholar

When no keyword matches, the wizard treats the query as multidisciplinary and suggests OpenAlex, Semantic Scholar and CrossRef. Searching the Literature explains how the wizard fits into a search.

OpenAlex

OpenAlex is a broad index of scholarly works across all disciplines, and it needs no key. If you set OPENALEX_MAILTO, or CROSSREF_MAILTO as a fallback, PRIMA Scholar sends that address with each request to join the OpenAlex polite pool.

• Open access: PRIMA Scholar uses OpenAlex’s own open access status for each work, with the open access link where OpenAlex has one.
• Dates: PRIMA Scholar sends both the exact date window and the year range to OpenAlex.
• Order: OpenAlex returns its most cited matches first by default, its newest when you sort by date and its best matches when you sort by relevance.
• Paper lookup: OpenAlex IDs such as W1234567890, or a full https://openalex.org/ address, go to OpenAlex.

PubMed

PubMed indexes biomedical and life sciences literature. It works without a key, and a PUBMED_API_KEY raises PRIMA Scholar’s PubMed limit from 3 to 10 requests a second.

• Open access: PubMed records do not say whether a paper is open access, so PRIMA Scholar marks every PubMed result as gated.
• Dates: PRIMA Scholar passes the date window or year range to PubMed as a publication date range.
• Order: PubMed lists the newest first when you sort by date, and uses its own order otherwise.
• Paper lookup: use a PubMed ID with its prefix, such as PMID:38123456.

Note

PRIMA Scholar marks every PubMed result as gated, so a search with open_access_only set to true drops all PubMed results. Europe PMC covers much of the same literature and reports open access status, so include it when open access matters.

arXiv

arXiv holds preprints in physics, mathematics, computer science, quantitative biology, finance and statistics. It needs no key, and PRIMA Scholar sends at most one request every three seconds.

• Open access: PRIMA Scholar marks every arXiv result as open access with full text available.
• Dates: PRIMA Scholar applies the window to the date each paper reached arXiv.
• Order: arXiv lists the newest submissions first when you sort by date, and uses its relevance order otherwise.
• Paper lookup: use an arXiv ID with its prefix, such as ARXIV:2301.12345.
• Citations: PRIMA Scholar records no journal for arXiv papers. Its APA7 citation names arXiv as the source, and the other five styles show no source.

Semantic Scholar

Semantic Scholar covers all disciplines and holds the citation graph that PRIMA Scholar uses to follow citations. A SEMANTIC_SCHOLAR_KEY is optional, and PRIMA Scholar sends it with each request. The built-in limit of 100 requests every five minutes stays the same with or without a key.

• Open access: PRIMA Scholar uses Semantic Scholar’s own open access status, with full text available when Semantic Scholar lists an open access PDF.
• Dates: PRIMA Scholar passes the date window or year range to Semantic Scholar.
• Order: Semantic Scholar uses its own order, since PRIMA Scholar sends it no date or citation sort.
• Paper lookup: Semantic Scholar paper IDs go here, and a DOI lookup tries Semantic Scholar first.
• Citation tracking: scholar_citations uses Semantic Scholar only, for both directions.

CrossRef

CrossRef holds DOI records with publication details for papers in every discipline, and it needs no key. If you set CROSSREF_MAILTO, PRIMA Scholar includes that address in each request to join the CrossRef polite pool.

• Open access: PRIMA Scholar marks a result as open access when its record carries a Creative Commons or open access licence link.
• Full text: CrossRef provides publication details only, never full text.
• Dates: PRIMA Scholar passes the date window or year range to CrossRef.
• Order: CrossRef uses its relevance order by default, lists the newest first when you sort by date and the most cited when you sort by citations.
• Paper lookup: a DOI lookup falls back to CrossRef when Semantic Scholar does not find the DOI.

CrossRef records often have no abstract, and some carry placeholder publication dates far in the future. To keep those placeholders out, PRIMA Scholar caps the window at today whenever you sort by date. With a date sort, CrossRef returns the newest records that match any word in the query, and on a broad query these can all fall on the last day of the window. A relevance sort avoids that for CrossRef.

CORE

CORE holds open access papers only. It is the one database that needs a key, and the key is free. Set it as CORE_API_KEY, as API Keys and Configuration explains.

• Open access: every CORE result is open access, with full text available when CORE holds the text or a download link.
• Dates: PRIMA Scholar sends only the year range to CORE and applies the finer date window afterwards on your computer.
• Order: CORE uses its own order.
• Full text: with a key, scholar_full_text looks the ID up as a CORE work and returns the text itself when CORE holds it. CORE is the only route that returns text, while every other route returns a link.
• Paper lookup: scholar_get_paper has no CORE route, so use the paper’s DOI instead.

Europe PMC

Europe PMC covers biomedical and life sciences research, and it needs no key.

• Open access: PRIMA Scholar uses Europe PMC’s own open access status, with full text available when the paper is open access and has a PMC ID.
• Open access filter: when you ask for open access papers only, PRIMA Scholar asks Europe PMC for open access papers too.
• Dates: PRIMA Scholar passes the date window as a first publication date range, or the year range as publication years.
• Order: Europe PMC lists the newest first when you sort by date, and uses its own order otherwise.
• Paper lookup: use a PMC ID such as PMC1234567.
• Full text: scholar_full_text returns the Europe PMC article page link for an open access paper that has a PMC ID, whether you give its PMC ID, PubMed ID or DOI.

ERIC

ERIC is a database of education research, and it needs no key.

• Search: PRIMA Scholar sends your whole query to ERIC as one exact phrase, matched in the title or the description. A long query therefore finds less in ERIC than in other databases, so keep ERIC queries short.
• Open access: PRIMA Scholar marks a result as open access when ERIC holds a full-text file, and links to that PDF.
• Dates: PRIMA Scholar sends only the year range to ERIC and applies the finer date window afterwards on your computer.
• Paper lookup: use an ERIC number such as EJ1234567 or ED123456.

bioRxiv/medRxiv

bioRxiv and medRxiv are preprint servers for biology and for medicine and health. PRIMA Scholar searches both together under the one name biorxiv, and neither needs a key.

• Search: PRIMA Scholar does not send your query to bioRxiv or medRxiv. Instead it reads the first page of preprints in the date range from each server. It keeps those whose title or abstract contains any space-separated term of three or more characters from your query, matched anywhere in the text and without regard to capitals.
• Dates: with no date set, the range runs from 1 January last year to 31 December this year.
• Open access: every result is a preprint, and PRIMA Scholar marks each one as open access.
• Full text: scholar_full_text returns the https://doi.org/ link for a DOI that starts with 10.1101/.
• Results: PRIMA Scholar lists bioRxiv matches before medRxiv matches and keeps only max_results in total, so medRxiv preprints appear only when bioRxiv returns fewer matches. To see medRxiv results, raise max_results or narrow the date window.

Note

PRIMA Scholar reads only the first page of preprints in the date range, so results from bioRxiv/medRxiv are a recent sample. Treat them as a starting point, and use a narrow date window for the closest match to a full search.

DBLP

DBLP is a bibliography of computer science publications, and like most of the databases here it needs no key.

• Open access: DBLP holds publication details only and has no open access information, so PRIMA Scholar marks every DBLP result as gated.
• Dates: PRIMA Scholar sends only the year range to DBLP and applies the finer date window afterwards on your computer.
• Paper lookup: scholar_get_paper has no DBLP route, so DBLP keys do not work there. Use the paper’s DOI instead.