Pricing
You pay each model maker's list price per token, and Veil adds nothing to it. Veil earns a published 5% fee when you add credit, and a published spread on the exchange rate when you pay in Toman.
Markup per token0%
Top-up fee5%
Exchange spread1.5%
Toman per dollar today—Top-ups paused: no fresh rate
How billing works
- Add credit in dollars. The quote lists the credit, the 5% fee, the rate and the Toman total, and holds for 15 minutes.
- Before a request starts, its worst case is reserved against your balance, so spending can never pass your credit or your caps.
- When the answer ends, the exact cost is settled from the model's price card and the rest of the reservation is released.
- Every request's cost lines, with the price card version and its source, are in the console.
What you're never charged for
- An answer that comes back empty, with no usable content.
- An attempt that failed over to another provider before the first byte. Only the provider that answered is billed.
- A request the gateway itself lost.
- Counting tokens, listing models, and looking up a key or a past request.
Credit never expires, and you can get it back
Unused credit is refunded in Toman at the lower of the rate you paid and today's rate, minus the bank transfer fee, newest credit first. The 5% fee isn't refunded. Refunds are paid by bank transfer.
Limits you set
Each key can have a spending cap per day, week or month, plus request, token and concurrency limits. Your organization can have a monthly budget. Alerts arrive at 50, 80 and 100% of any cap, and when your balance falls below the threshold you choose.