Academic Paper Search — Crossref DOIs, Authors, Citations

Search the Crossref registry — the metadata backbone of scholarly publishing, roughly 160 million registered works — by topic, author, journal, funder, date and licence. Get back DOIs, authors with their ORCID identifiers and affiliations, citation counts, abstracts and open-access links.

What comes back

Searching solid state battery in titles, 2025 onward:

Title	Journal	Cited by	Open
Advanced Solid-State Lithium-Oxygen Battery	—	—	yes
Toward a full solid state aluminum battery	—	—	yes
Phase Diagrams Enable Solid-State Battery Design	—	—	yes

Each work carries its DOI, publication date, volume and pages, every author with ORCID and affiliation, citation and reference counts, subjects, funders, licence URLs, direct full-text links where the publisher provides them, and a relevance score.

Choosing where to search

Title is the right default for a topic. It is the most precise field and keeps the results on subject.

Everything searches titles, authors and journal names together. Broader, and useful when you are not sure how a field names a thing — but a word like state or battery will pull in unrelated work, so expect to skim.

Author finds a researcher's output; combine it with withOrcidOnly to follow one person across institutions and name spellings rather than everyone who shares a surname.

Journal shows what a specific publication has printed.

A note on sorting

Relevance is the default and usually right. Sorting by citation count reorders everything Crossref considered a match, so with a broad query it will surface heavily-cited work that is only loosely on topic — the most-cited genomics paper of 2025 outranks a precise battery paper on almost any broad query. Pair citation sorting with a narrow title search, or keep relevance.

Sorting by publication date is how you watch a field week by week.

For building a corpus

Three filters matter if you are feeding a model or building a search index:

withAbstractOnly keeps only works whose publisher deposited an abstract — a title alone carries little signal. openAccessOnly keeps works you can actually read in full. workTypes set to posted-content gives you preprints, where new results appear months before formal publication.

Results page reliably in the thousands, so a topic corpus is one run.

Who uses this

Researchers and librarians running systematic reviews. R&D and competitive-intelligence teams watching a technical field. Anyone building a RAG index or a literature dataset who needs citable metadata rather than scraped pages. Funders and institutions tracing what their money produced, via the funder DOI filter.

Source and licence

Crossref, through its public REST API and the polite pool it recommends for identified callers. Crossref metadata is openly available and free to reuse; the Actor sends a contact address with every request as the API asks.

Abstracts and full text belong to their publishers. This Actor returns the abstract where the publisher deposited it with Crossref and a link where they provide one; it does not fetch article text from behind a paywall.

Frequently asked

Why is an abstract missing? Because the publisher never deposited one. Coverage varies enormously by field — strong in life sciences, patchy in the humanities. withAbstractOnly filters to what exists.

Why is a paper I know of absent? Crossref indexes registered DOIs. A work without a DOI, or one registered with DataCite instead, will not appear.

Are citation counts the same as Google Scholar's? No. Crossref counts citations from works that deposited their references with Crossref, so the figure is lower and more conservative. It is consistent and auditable, which is what matters for comparison.

Can I fetch specific papers instead of searching? Yes — put DOIs in the dois field, with or without the https://doi.org/ prefix, and each is fetched directly.