Clear Skies became the first company to offer a commercial papermill detection service in 2022.
That service was our Papermill Alarm Early Warning API delivered via RapidAPI. Your developers can sign up to the RapidAPI service here and start using it today.
The Papermill Alarm itself is built on confidential data sources and is available exclusively by direct agreement with Clear Skies. Contact us for details.
The Papermill Alarm API detects rare signals in the research literature
How it works
- Users provide data. We analyse and model that data perpetually and return the Papermill Alarm: the best research fraud detection system in the world.
- We recommend using the Papermill Alarm wherever possible, but where users are unable to provide data, we analyse publicly available data and return the Early Warning API, which is the best-possible papermill detection system built on that data.
There are 2 versions of the API
- Early Warning API. This tool is offered under license through RapidAPI, STM Solutions, Morressier and other intermediaries. We use publicly-available data from OpenAlex, Crossref, ORCID, PubMed and other data sources to model known cases of organised research fraud. The API will issue an alert when a new submission matches patterns observed in known-milled papers. The API is intended to provide an early warning on new-submissions. It is not intended to be an investigation tool on its own. However it does also return non-sensitive data which can support editorial investigations concerning:
- author-histories (based on ORCID),
- referencing practices,
- and tortured phrases (using the list of tortured phrases provided by the Problematic Paper Screener).
- Papermill Alarm. This is based on “private” data which is data received under direct data sharing agreements with our customers. We never share data received from customers with any third party. Instead, we use it to build analytical models which serve the private API. The private API performs the same tasks as the Early Warning API, with improved accuracy and full data output. It also makes sensitive output from our AI and network-analyses available.
| Early Warning API – available from RapidAPI – also sublicensed by our partners | Papermill Alarm – available with direct agreement | |
| Early warnings Article-level alerts at the moment of submission, intended for editorial workflows | ✔ | ✔ |
| Output limited to non-sensitive data | ✔ | |
| Alerts based on publicly available data | ✔ | ✔ |
| Alerts based on all available data | ✔ | |
| Suitable for misconduct investigations | ✔ | |
| Full article-level insight including AI and network analysis through our dashboard web app | ✔ | |
| Integration with your peer-review platform | ✔ |
Article Tracker
We also offer the Article Tracker API. This API was the first tool to detect the duplicate submission practices commonly used by papermills. It detects duplicate submissions cheaply and at scale using the Crossref Public API.
How it works
- The API searches Crossref for articles with similar metadata. It then retrieves candidate matches. This allows you to
- Track rejected articles
- Identify duplicate submissions (a common feature of papermilling)
- Track abandoned articles (also a common feature of papermilling)
- With further data analysis, this tool can show you patterns in how papermills target journals, including your own journals and those of your competitors.