Semantic Scholar Academic Graph (S2AG)
Allen Institute for AI (Ai2)
214 million papers and 2.49 billion citations, with AI-extracted influential citations and paper embeddings.
The open graph behind the Semantic Scholar search engine, available through an API and as bulk datasets. Records cover papers, authors, citations with context snippets, abstracts, and paper embeddings; S2ORC adds parsed full text for open access papers. It is heavily used in NLP and science-of-science research.
Access
- Cost
- FreeFree; most API endpoints work without a key, but bulk datasets and higher rate limits need a free API key.
- License
- Provider's own termsSemantic Scholar API License Agreement
- Sign-up
- Free account
- How you get it
- API and Bulk downloadJSON, API
What's in it
- One row is
- One record per paper, author, or citation
- Covers
- Worldwide research output, strongest in computer science and biomedicine
- Years
- Historical literature to present
- Updated
- Weekly
- Size
- 214 million papers, 2.49 billion citations, 79 million authors
Quality
Large and actively maintained with useful derived fields, but entity resolution is automated.
- Author disambiguation errors are common
- Field coverage is uneven outside computing and medicine
Connects to
- OpenAlex: Alternative open graph to cross-check citation counts
- arXiv Metadata (OAI-PMH): Preprints matched to their S2 records
Questions it could answer
- Which papers are cited as influential rather than in passing?
- How quickly do citations to a paper peak, by field?