Plans and billing
You pay for two things, both through AWS:
- Routing, to AI Cloud Gateway through AWS Marketplace: your plan, plus requests beyond its monthly allowance.
- Bedrock tokens, to AWS in the account you connected, at your Bedrock prices. AI Cloud Gateway adds no markup and does not resell tokens.
Plans
| Plan | Monthly | Annual | Routed requests per month | Beyond the allowance | Support |
|---|---|---|---|---|---|
| Starter | $19 | $190 | 50,000 | $0.30 per 1,000 | Email, next business day |
| Team | $139 | $1,390 | 1,000,000 | $0.25 per 1,000 | Faster email response |
| Business | $649 | $6,490 | 5,000,000 | $0.22 per 1,000 | Priority |
Every plan routes the same way and covers one connected AWS account. The AWS Marketplace listing's pricing tab is the final reference.
How usage is billed
- Only answered requests count. Refused calls (
400,401,403,404, or every model failing), including calls refused by a budget, andGET /modelsare free. - A response your guardrail blocked (
finish_reasoncontent_filter) is an answered request. - The allowance resets at 00:00 UTC on the first of each month. Unused requests do not carry over.
- The dashboard's Usage page shows this month's allowance used. Charges appear in the AWS Billing console of the subscribing account.
Budgets are not the allowance
Going past the allowance stops nothing; extra requests are billed per 1,000. A budget is an optional limit you set, and requests are refused once it is reached. To keep routing charges close to your plan's fee, set a monthly request limit equal to your allowance.
Change or cancel
- Upgrade any time in AWS Marketplace; the larger allowance applies at once.
- Downgrade takes effect at renewal.
- Cancel by turning off auto-renewal; the plan runs to the end of its term.
Full guide: billing.md on GitHub