Research Compass uses 220+ million academic
publications to answer questions with Retrieval
Augmented Generation (RAG), re-ranking, and a prompt
designed to surface differences in the sources. Some
of the publications are open access and some are
limited to the abstract. Our data pipeline starts
with
PubMed,
Semantic Scholar,
CORE,
PubMed Central, and
arXiv.
We supplement this data with our own transformations
and additions, including data from publishers,
organizations like
Crossref
(and their
Retraction Watch
database), lists of predatory journals, and journal
rankings. We use AI classification tools to annotate
these abstracts with metadata about referenced
studies, surveys, datasets, experiment design, and
other ways to evaluate scientific methods.
We are not affiliated with any of these
organizations.