Skip to main content

Semantic map controls

Semantic map controls help you explore the project-level landscape across sources, notes, chats, and concepts. Unlike a source-level knowledge map, a semantic map shows how project items relate to each other.

Core elements

ElementMeaning
Point or itemA source, note, chat, or project artifact in the projection.
ClusterA group of semantically related items.
DensityA region where many related items occupy similar space.
LabelA generated description of a region, theme, or item.
Angle promptA user-provided perspective that changes the projection.
SearchA way to find a concept or item on the map.

Spatial interpretation

Items close together are semantically related in the current projection. Distance is a reading aid, not a mathematical proof of importance or causality.

Outliers can be useful. They may represent unrelated material, a unique method, a different domain, or a source that needs review.

Angle prompts

An angle prompt changes the lens for the map. For example, the same project can be projected around "methods", "limitations", "clinical relevance", or "commercial opportunity".

Use a concrete angle when the default map is too broad. Avoid vague angles like "interesting things" unless you are intentionally exploring.

Opening items

Selecting an item should move you from the visual overview to the underlying material. The map is most useful when it helps you decide what to read, compare, or ask about next.

Semantic map versus knowledge map

MapScopeBest for
Semantic mapProject-level landscape across items.Finding clusters, gaps, and relationships across the workspace.
Knowledge mapSource-level or topic-level structure.Understanding concepts and relationships inside a source or focused source set.

Reading caution

Generated clusters and labels can be useful but imperfect. Open the underlying items before using a cluster label as a research conclusion.