Knowledge graph

The graph draws your links as a network. It is for the questions a list cannot answer — what is this connected to, and what is that connected to.

Two modes

Explore starts from one record and walks outward. Pick a seed with Choose entity, or arrive here from a record's "View in graph" action.

Whole graph shows everything you have linked, in one picture. It is capped so it stays legible rather than becoming a hairball.

Moving around

Click a node to select it — the panel shows what it is and its relationships. From there:

• Center here re-roots the exploration on that node, walking outward from the new centre;
• Open record leaves the graph for the record itself.

Re-centring repeatedly is the intended way to travel: each step reveals the next neighbourhood rather than showing you everything at once.

Filtering by type

The Types control limits which kinds of record are drawn. Turning off the type that dominates your data — usually tasks — is often what makes the structure visible.

Email nodes

One node type is not a link you made: email addresses, derived by matching addresses across records. They connect things that share a correspondent.

That is why the graph sometimes shows a connection you do not remember creating. It is not wrong — it is the one inference Kancil makes, and it is usually the interesting one. Email nodes route to Mail; they are not records you can link to by hand.

What it is not

The graph is a reading tool. You cannot create links by dragging nodes together — links are made on the records themselves, through the Related panel. The graph shows you what is there.