Prepaid, per-token, no subscription, no minimum, no card. Every request is charged at the official published token price × 0.30. Nothing expires, and when the balance reaches zero, requests simply stop.
| You pay | Official-priced usage | Typical use |
|---|---|---|
| $5 | ≈ $16.67 | a few evenings of coding |
| $10 | ≈ $33.33 | a light week of agent work |
| $20 | ≈ $66.67 | heavy Claude Code / Codex CLI use |
| $50 | ≈ $166.67 | a small team for a month |
| Model class | Official list (per M tokens) | You pay |
|---|---|---|
| Opus-class | $5 input / $25 output | $1.50 / $7.50 |
| Sonnet-class | $3 input / $15 output | $0.90 / $4.50 |
| Codex line (gpt-5.5) | official rate | 30% of it |
Official list prices are set by the model vendors and can change; the 0.30 multiplier is what stays fixed. Cached input tokens are billed at the official cache rate × 0.30, and keeping one long session open is dramatically cheaper than restarting it often.
| Consumer plan | Official API | MadeMore API | |
|---|---|---|---|
| Commitment | $20–200 / month | none | none |
| Unused credit | lost each month | — | never expires |
| Can you overspend? | no | yes | no — hard stop at zero |
| Card required | yes | yes | no |
| Payment | card | card | USDC on Base, USDT on Polygon / BNB Chain |
| SLA | vendor | vendor | none — best effort, stated openly |
We buy capacity in bulk at a wholesale rate and resell it through this gateway. You are paying for the same models at a third of the price, from a third party rather than the vendor — that trade-off is the whole product.
No. The smallest top-up is $1 and there is no recurring charge of any kind.
Credits are matched automatically by the unique amount on each order, usually within a minute. If a transfer lands late or on the wrong network, email the transaction hash and we credit it manually.
Unused top-ups: refundable on request within 7 days, returned in the same asset and network. Consumed balance is not refundable, because it has already been spent upstream.
Ask by email with your use case. For larger prepayments we can discuss volume terms — but be aware there is no SLA behind this service.