Ask the whole tree

A tree can grow large — dozens of branches, each with its own conclusion. Tree insights (in the chat header) lets you ask one question and get a synthesis across every branch at once, instead of opening them one by one.

The same panel shows the token savings: how much less context the tree sends the model than the same conversation would as one linear thread — with a live count of branches and messages. If the number surprises you, the token-economics explainer is why it’s real.

On the free plan, asking across a whole tree is capped by tree size; a paid plan removes the cap.