Knowledge Graph

Visualization

Inspect how saved links relate through shared tags, source overlap, and save-time proximity.

Overview

The Knowledge Graph view is a fullscreen 3D network for your saved archive. Each bookmark becomes a node, and links are drawn when two items are correlated by shared tags, the same source domain, or being saved near each other in time. Open it from the sidebar or visit /dashboard/connection.

Unlike the knowledge graph, this page favors explainable, user-visible signals over embedding similarity so you can understand why two links sit near each other without reading a ranking formula.

Correlation Signals

Shared Tags

Tags are the strongest signal. Each shared tag adds weight to an edge, so links that repeatedly overlap in topic cluster together more tightly.

Same Source Domain

Links from the same publication or origin get an additional relationship boost. This helps reveal source-heavy rabbit holes inside the archive.

Recent Save Windows

Nearby save timestamps create lighter correlations, surfacing bursts of research activity even when tags are sparse.

Interactions

DragRotate the graph and reposition nodes directly on the canvas.

ScrollZoom the camera to inspect dense clusters or zoom back out for the archive-wide view.

ClickSelect a bookmark to focus its neighborhood and open the shared inspector from the side panel.

FilterSearch by title, domain, or tag, and switch the signal filter to isolate a specific type of relationship.

Tip: The view is most useful when bookmarks have tags or summaries. Processing items stay out of the graph until they have enough metadata to contribute meaningful correlations.