Cito is built entirely on openly licensed scholarly data. This page lists every source, its license, and what that means for how you use API results.
Sources and licenses
- Semantic Scholar corpus (metadata, abstracts, citations)ODC-BY 1.0Commercial use and redistribution allowed; credit Semantic Scholar.
- Semantic Scholar SPECTER2 embeddingsApache-2.0Commercial use allowed; retain the license notice.
- OpenAlex (open-access links, abstract backfill)CC0Public domain; no attribution required (we credit them anyway).
- SPECTER2 model (allenai/specter2)Apache-2.0Commercial use allowed.
What this means for you
- You may use Cito results in commercial products, research, and publications.
- When you republish results (in an app, a dataset, a paper), keep the
attributionstring that every API response carries: "Data from Semantic Scholar". That credit is the one obligation ODC-BY imposes downstream. - PDF links point to the copy at its origin (publisher or repository). Cito never proxies or re-hosts full text, and what you may do with a downloaded PDF is governed by that origin's own terms, not by Cito.
Credits
Cito exists because the Allen Institute for AI and OurResearch publish Semantic Scholar and OpenAlex under open licenses. If your use case outgrows Cito, both are excellent projects to support directly.
The dataset licenses ask products built on this data to cite the papers describing it. If Cito contributes to published work, please cite Kinney et al., The Semantic Scholar Open Data Platform (2023) for the corpus, and Singh et al., SciRepEval (2022) for the SPECTER2 embeddings.