Pricing
Two multiplications, applied to every request on every model. Nothing else to learn.
you = direct × 0.66 · funder = you × 0.75 · platform = you − funder
Ratios live in server env (UNBURN_SPENDER_RATE / UNBURN_FUNDER_SHARE), not in anyone's head.
Every model with funder supply right now. Provider list price struck through, your price in signal. Refreshed from upstream daily.
Direct cost comes from per-token provider rates matched by canonical model name. Custom names on OpenAI-compatible endpoints can't be priced — name funder models canonically. Settlement rounds to cents at the payout boundary.