Waiting for calls. Ask Dora something in the other window.
LLM transactions
LLM cost = (input + output tokens) × a flat demo rate. Each Dora reply can make several model calls (one per round of tool use).
Compute · Maritime
Each citizen gets their own always-on Dora on Maritime, billed flat per agent per month. “Busy” is the time Dora actually worked on a reply, priced at the same rate, to compare one reply's compute with its LLM cost.