Semantic Scholar Papers Search

Semantic Scholar compared with Google Scholar, Scopus and Web of Science
Semantic Scholar	Google Scholar	Scopus	Web of Science
Type and operator: Semantic Scholar is a free, AI powered research tool for scientific literature, built by Ai2 (the Allen Institute for AI).	Type and operator: Google Scholar is a search service that indexes peer-reviewed papers, theses, preprints, abstracts and technical reports from publishers.	Type and operator: Scopus is an abstract and citation database from Elsevier.	Type and operator: Web of Science is a citation index from Clarivate, whose Core Collection has six component indexes including Science Citation Index Expanded and Emerging Sources.
Scale and coverage: The site searches about 239 million papers from all fields of science.	Scale and coverage: Coverage spans all research disciplines. Inclusion depends on publisher sites meeting technical guidelines, for example a searchable PDF of limited size and one file per article.	Scale and coverage: Provider figures list more than 21,000 peer reviewed journal titles from over 7,000 publishers, plus conference papers, book titles and preprints. The sources skew toward English language and STEM titles.	Scale and coverage: Provider figures list more than 22,000 journals in the Core Collection. Inclusion criteria are described as more selective, and the Science Citation Index dates from 1964.
Access and cost model: Searching is free and works without signing in. The API datasets are described as free and open resources.	Access and cost model: Google does not provide usage statistics to publishers, and web server logs are the source.	Access and cost model: An institutional subscription is required, and there is no free public search interface.	Access and cost model: Core Collection access requires an institutional subscription. A free Master Journal List shows whether a journal is currently indexed.
Programmatic access: A public API offers the Academic Graph (authors, papers, citations, venues, SPECTER2 embeddings), recommendations and downloadable datasets. Most endpoints work without a key, and the introductory limit for an API key is 1 request per second.
Citation analysis and metrics: Author pages and citation lists trace how an idea spread between papers, and the Semantic Reader is an augmented reading interface in beta.	Citation analysis and metrics: Citation counts factor into ranking, and versions of a work are grouped so that their citations count toward one work.	Citation analysis and metrics: Scopus calculates CiteScore, with percentile and quartile rankings per subject category, for actively indexed sources.	Citation analysis and metrics: Journal Impact Factor is published annually in Journal Citation Reports, and only journals in SCIE or SSCI receive one.

Semantic Scholar compared with Google Scholar, Scopus and Web of Science

Semantic Scholar

Google Scholar

Scopus

Web of Science

Type and operator: Semantic Scholar is a free, AI powered research tool for scientific literature, built by Ai2 (the Allen Institute for AI).

Type and operator: Google Scholar is a search service that indexes peer-reviewed papers, theses, preprints, abstracts and technical reports from publishers.

Type and operator: Scopus is an abstract and citation database from Elsevier.

Type and operator: Web of Science is a citation index from Clarivate, whose Core Collection has six component indexes including Science Citation Index Expanded and Emerging Sources.

Scale and coverage: The site searches about 239 million papers from all fields of science.

Scale and coverage: Coverage spans all research disciplines. Inclusion depends on publisher sites meeting technical guidelines, for example a searchable PDF of limited size and one file per article.

Scale and coverage: Provider figures list more than 21,000 peer reviewed journal titles from over 7,000 publishers, plus conference papers, book titles and preprints. The sources skew toward English language and STEM titles.

Scale and coverage: Provider figures list more than 22,000 journals in the Core Collection. Inclusion criteria are described as more selective, and the Science Citation Index dates from 1964.

Access and cost model: Searching is free and works without signing in. The API datasets are described as free and open resources.

Access and cost model: Google does not provide usage statistics to publishers, and web server logs are the source.

Access and cost model: An institutional subscription is required, and there is no free public search interface.

Access and cost model: Core Collection access requires an institutional subscription. A free Master Journal List shows whether a journal is currently indexed.

Programmatic access: A public API offers the Academic Graph (authors, papers, citations, venues, SPECTER2 embeddings), recommendations and downloadable datasets. Most endpoints work without a key, and the introductory limit for an API key is 1 request per second.

Citation analysis and metrics: Author pages and citation lists trace how an idea spread between papers, and the Semantic Reader is an augmented reading interface in beta.

Citation analysis and metrics: Citation counts factor into ranking, and versions of a work are grouped so that their citations count toward one work.

Citation analysis and metrics: Scopus calculates CiteScore, with percentile and quartile rankings per subject category, for actively indexed sources.

Citation analysis and metrics: Journal Impact Factor is published annually in Journal Citation Reports, and only journals in SCIE or SSCI receive one.