DocsCitation graph
Read with AI

Citation graph

Every paper detail page surfaces a List view and a Graph view of the work's citation neighbourhood. The graph is interactive: click any node to expand it.

What's in the graph

  • Center — the paper you opened. Cream-coloured.
  • Cited by — papers that cite this one. Purple.
  • References — papers this one cites. Slate blue.
  • Related — conceptually adjacent papers. Tan, drawn dashed.

Node size scales with citation count (square-root scaled so very high-impact papers don't dwarf everything else). Arrows show the direction of citation.

Click vs shift-click

  • Click a non-center node to expand it: Kvire fetches that paper's top neighbours and merges them into the same graph. The node gets a subtle ring once expanded.
  • Shift+click to navigate to that paper's detail page instead.

Expand-on-click lets you build a multi-hop neighbourhood without leaving the page. Each expansion adds at most 5 cited + 5 citing + 3 related to keep the canvas readable.

Where the data comes from

Citation edges come from OpenAlex — they're pre-computed, not synthesized by Kvire. So the same edges you see in OpenAlex's catalog are what we render.

Limitations

  • Books don't have a citation graph (Open Library doesn't track them that way).
  • Very recent papers may have empty “cited by” — the citation network is sparse before papers accumulate references in the catalog.
  • The graph is positional, not chronological. Use the labels (visible at higher zoom) to orient yourself.