A per-call cap and a total budget are easy to set and hard to reason about. The recorded run shows why: a $0.0025 budget bought $0.002 of evidence and left the answer $0.051 short. This page prices that trade for your facts and your caps, with the same solver. Your own facts are priced over the live Kite catalog, read today over plain HTTPS with no key. The recorded run is priced over the catalog of the day it ran, because that is the catalog its prices belong to. No account, no session, nothing moves.
Working out the price…
The solver is exact, not clever. It enumerates the subsets of your facts, costs the cheapest catalog endpoint that covers each one, takes the basket that fits inside your caps, and reports what is left over. There is no model in the loop and no cut-off, so the same inputs always give the same answer.
The catalog is the live Kite service catalog, filtered to Base mainnet USDC.
Your facts are priced over it as it stands today, read over plain HTTPS with
no key and no account. The recorded run is priced over the catalog of the day
it ran, which is the catalog its prices belong to, and that answer is the
number the run measured. Prices belong to the merchants. The same answer is
available as /api/limit-cost and as the
price_my_limit tool in the @oxunify/capsule-mcp package.