To use Connected Papers, begin by inputting a single "seed paper" using its title, DOI, PMID, or relevant keywords. The platform then processes this input, analyzing tens of thousands of academic papers, typically around 50,000, from the extensive Semantic Scholar Paper Corpus. Its advanced algorithm identifies dozens of papers with the strongest conceptual connections to your original seed paper. Instead of a linear list, Connected Papers generates an interactive graph. In this graph, papers are arranged based on a sophisticated similarity metric derived from co-citation and bibliographic coupling. This means that articles sharing significant overlap in their citations and references are considered related and are visually clustered together, even if they don't directly cite one another. The system employs a Force Directed Graph algorithm to effectively distribute these papers, ensuring that similar works are grouped closely while less similar ones are pushed apart. Each circle, or node, on the graph represents an academic paper. The size of a node visually communicates its total number of citations, while its color indicates the publication year, with darker shades signifying more recent publications. The closer two nodes are, the stronger their connection and similarity. This dynamic, map-like overview of a research area helps users quickly grasp the structure and evolution of a field, revealing hidden connections. For deeper insight, simply click on any node in the graph or an item in the detailed list view to display comprehensive information about that specific paper. You can also generate an entirely new graph based on any related paper you discover, allowing for continuous and in-depth exploration of interconnected research.