Pricing
One fee, taken once, where you can see it. Provider list price passes straight through, the number in the catalog is the number you are charged.
Platform fee
2.00%
taken once, when you add credit
- No per-token markup
- No subscription
- No minimum monthly spend
- Credit does not expire
Per request
0%
markup on provider list price
Every response returns its exact cost as a decimal string, broken down by input, output, cache read and cache write. You never have to take our word for the arithmetic.
Why the fee sits at deposit
A per-token markup can be folded into the sticker price and then advertised as “0% platform fee”, which is technically defensible and practically misleading. Charging once at deposit makes the fee a single visible number and leaves model prices identical to the provider's own published rates, so you can check them yourself against the provider's pricing page.
| Model | Input / 1M | Output / 1M | Routes |
|---|---|---|---|
| openai/gpt-4.1-nano | $0.10 | $0.40 | 1 |
| openai/gpt-4.1-mini | $0.40 | $1.60 | 1 |
| openai/o4-mini | $1.10 | $4.40 | 1 |
| openai/gpt-4.1 | $2.00 | $8.00 | 1 |
Adding credit
- Chain
- HyperEVM mainnet (999)
- Token
- USDC (6 decimals)
- Minimum deposit
- $5.00
- Confirmations
- 1 (HyperBFT commits are final)
The payments contract is not deployed in this environment, so the deposit flow is unavailable. Deployment steps are in docs/PAYMENTS.md.
Model prices are the providers' own published rates, recorded in src/lib/providers/manifests.ts with the date they were last verified.