TRANSPARENT PAY AS YOU GO

Know the maximum before every request.

Prepaid billing reserves the worst-case amount first, charges only delivered usage and returns the unused reserve automatically.

GPT-4o Chat

$5 / $20
per 1M input / output tokens
  • OpenAI-compatible Chat Completions endpoint
  • Reserve calculated from prompt tokens and max_tokens
  • Final charge calculated from upstream usage

Managed Web Extraction

$0.001725
current standard price per successfully delivered request
  • Approved public HTTPS targets only
  • Managed upstream and optional geographic routing
  • Failed upstream delivery releases reserved funds

Exact settlement, not an estimate

The financial sequence is enforced before and after every GPT-4o request.

1 · RESERVEInput prompt plus the requested maximum output is priced before the upstream call.
2 · EXECUTEIf the available balance cannot cover the reserve, the API returns HTTP 402 and does not call the provider.
3 · SETTLEActual prompt and completion usage determines the final charge; the difference returns to the balance.
Worked example: 1,500 prompt tokens and max_tokens=2,000 reserve $0.047500. If actual completion is 800 tokens, the final customer charge is $0.023500 and $0.024000 is returned.
Prices shown in USD. Accounts must be funded before provider-backed usage. Standard prices may be updated prospectively; completed usage is always settled using the price recorded for that transaction.

Start with a controlled prepaid budget.

Create an account and top up securely, or request a non-sensitive cost benchmark for your workload.

Create accountRequest benchmark