Knowledge Graph

A knowledge graph contains entities, relations, and evidence. Entities may represent characters, locations, organizations, items, rules, or events; relations should express verifiable connections.

1. Choose a source: manuscript, reference, outline, a combination, the current chapter, or another selected range.
2. Start extraction and follow batch progress.
3. Review nodes, edges, and conflict candidates.
4. Merge aliases and duplicate entities, and correct wrong types.
5. Save a graph version and reopen the workspace to verify it.

If loading stalls, wait for extraction or saving to finish, leave and return, reload the project, and inspect the graph version and error. Do not repeatedly start extraction while loading. 2D suits editing; 3D helps inspect large structures. Layout changes do not alter graph data.