Master Connected Papers with our step-by-step tutorial, detailed feature walkthrough, and expert tips.
Explore the key features that make Connected Papers powerful for research agents workflows.
Force-directed graph visualizations where node proximity indicates conceptual similarity (via co-citation and bibliographic coupling), node size reflects citation impact, and color encodes publication recency. The algorithm processes a seed paper's citation context to identify the most conceptually related works and arranges them spatially so that tightly related papers cluster together while loosely related ones sit further apart. This visual encoding allows researchers to immediately grasp the structure of a research area, spot dominant clusters, identify bridge papers connecting subfields, and detect outlier works that might represent novel or underexplored approaches.
Researchers exploring unfamiliar domains can instantly visualize the research landscape, identifying key papers, research clusters, and potential gaps without reading hundreds of abstracts
Three distinct visualization modes — similarity graphs, prior works graphs, and derivative works graphs — separate intellectual influences from subsequent impact. This creates a clear temporal narrative showing how ideas evolved from foundational theories to current applications. The prior works view traces backward through citation chains to reveal the theoretical and empirical foundations that shaped a paper's contributions, while the derivative works view maps forward to show how a paper's ideas have been extended, applied, challenged, or refined by subsequent research. Together these modes provide a complete intellectual genealogy for any indexed paper.
Graduate students building comprehensive literature reviews can trace the complete intellectual genealogy of their research area, from foundational theories to cutting-edge developments
Premium feature on the $36/year Academic plan that allows seeding graphs with multiple papers simultaneously, identifying research at the intersection of different fields, methodologies, or theoretical frameworks. Particularly useful for interdisciplinary teams, this feature generates visualizations that highlight papers bridging disparate domains — for example, connecting computational neuroscience with clinical psychiatry, or linking machine learning methods with environmental science applications. The resulting graphs reveal synthesis opportunities and collaboration possibilities that single-origin searches would miss entirely.
Interdisciplinary researchers working at the boundary of AI and healthcare can input key papers from both domains to discover bridges and synthesis opportunities
Built on Semantic Scholar's 200M+ paper corpus with automated updates from arXiv, PubMed, IEEE, ACM, and major publishers. Provides broader coverage than tools relying on narrower citation indexes, with new papers typically indexed within days to weeks of publication. The integration leverages Semantic Scholar's natural language processing capabilities for metadata extraction and paper matching, ensuring high-quality citation linking even for papers from diverse sources. This extensive corpus forms the foundation for Connected Papers' graph algorithms, enabling comprehensive coverage across STEM disciplines and growing representation in social sciences.
Systematic review teams need confidence that their search captures papers across journals, conferences, and preprint servers without manually querying multiple databases
Every generated graph receives a unique shareable URL for collaboration, and citation lists can be exported to BibTeX, RIS, and other reference manager formats. Integrates with Zotero, Mendeley, and EndNote workflows without requiring plugins or manual data entry. Researchers can share graph URLs with collaborators, advisors, or students to communicate their understanding of a research landscape visually. The export functionality allows seamless transition from discovery to citation management, enabling researchers to move selected papers directly into their reference libraries for use in manuscript preparation.
Research labs can collectively map emerging areas, share discovery insights, and maintain team knowledge bases while integrating with existing bibliography workflows
Traditional databases show direct citation relationships — papers that explicitly cite each other in their reference lists. Connected Papers uses co-citation analysis and bibliographic coupling algorithms to identify conceptual similarity between papers, even when they do not directly cite each other. This means it can surface relevant work that shares intellectual foundations with your seed paper but might never appear in a traditional forward or backward citation search. The visual graph format also communicates the structure of a research field at a glance, showing clusters, bridges, and outliers that would require hours of manual analysis to identify through linear database results.
No — Connected Papers excels as a discovery and mapping tool but should complement, not replace, systematic review protocols like PRISMA. Use it to identify key papers and research clusters rapidly, then validate comprehensiveness through traditional database searches with documented search strategies. Connected Papers is particularly valuable in the scoping phase of a systematic review, where it can help researchers understand the landscape, refine search terms, and identify relevant MeSH headings or keywords before conducting the formal protocol-driven search across multiple databases.
Coverage is strongest in STEM fields including computer science, biomedicine, physics, chemistry, mathematics, and engineering, drawing from Semantic Scholar's 200M+ paper corpus that aggregates from arXiv, PubMed, IEEE, ACM, and major academic publishers. Social sciences such as psychology, economics, and political science are reasonably well represented. Coverage is weaker for humanities disciplines like philosophy, literature, and history, as well as for non-English language publications and regional journals not indexed by major international databases. Researchers in underrepresented fields should treat Connected Papers as one discovery tool among several rather than a comprehensive source.
For actively researching graduate students, almost certainly yes. The free plan's 5 monthly graphs are typically consumed within 2-3 days during literature review phases of dissertation work. At $36 annually ($3/month), the Academic plan provides unlimited graphs, multi-origin graph creation for interdisciplinary exploration, priority processing for faster results, and advanced filtering options. Compared to the time cost of manual literature searching — often dozens of hours per review cycle — the subscription pays for itself after a single productive session. It is particularly valuable during proposal writing, comprehensive exam preparation, and the literature review chapter of a dissertation.
Connected Papers builds on Semantic Scholar's regularly updated corpus, which ingests papers from major publishers, preprint servers (arXiv, bioRxiv, medRxiv), and conference proceedings. New papers typically appear within days to weeks of publication or preprint posting, though the exact latency varies by source. However, very recent papers with few citations will produce sparse graph visualizations because the co-citation and bibliographic coupling signals need time to accumulate. For cutting-edge preprints, researchers should combine Connected Papers with direct preprint server monitoring and citation alert services for the most complete coverage.
Now that you know how to use Connected Papers, it's time to put this knowledge into practice.
Sign up and follow the tutorial steps
Check pros, cons, and user feedback
See how it stacks against alternatives
Follow our tutorial and master this powerful research agents tool in minutes.
Tutorial updated March 2026