Free
$0 / month
1,000 requests / month
No card required. Stops at the included limit.
No usage-based overage
Start freePricing
Start free. Scale by request volume.
AI gateway and LLM observability in every plan. Each account gets 1,000 base requests per UTC calendar month, shared across all projects. Paid subscriptions add their allowance to that base.
Starter
$9.99 / month
11,000 requests / month
Includes the shared free allowance.
$1.50 per extra 1,000 requests
Create an accountPro
$49.99 / month
101,000 requests / month
Includes the shared free allowance.
$0.75 per extra 1,000 requests
Create an accountPremium
$199.99 / month
1,001,000 requests / month
Includes the shared free allowance.
$0.30 per extra 1,000 requests
Create an accountBusiness
$1,499.99 / month
10,001,000 requests / month
Includes the shared free allowance.
$0.22 per extra 1,000 requests
Create an accountEnterprise
Custom
Custom volume
Discuss volume and deployment requirements.
Custom agreement
Contact usWhat counts as a request?
Gateway usage and accepted external metrics share the account allowance. Accepted external events count even when the reported provider call failed. Retry an external event with the same ID and payload to avoid another charge. Gateway rejections can appear in analytics without consuming quota.
What is included?
Request, token, cost and latency analytics; client and session reporting; gateway key protection, routing, limits, webhooks and prompt releases. Gateway controls apply to routed calls. Analytics-only integrations cannot enforce controls on direct provider calls.
Provider charges and tax
OpenAI, Anthropic, Google and other providers bill their API usage separately. HyperProxy plan prices are in USD and exclude VAT / sales tax, added according to your location. AI spend charts estimate provider costs and are separate from your HyperProxy invoice.
How usage-based billing works
It is enabled by default for paid plans. Extra requests are rounded up to blocks of 1,000 and billed on the next invoice at the applicable rate. You can turn it off in Account to stop at the included limit. When enabled, the account safety limit is 2× the total included allowance.
Starter example: one Starter subscription provides 10,000 + 1,000 base requests = 11,000 included. At 12,100 requests, the 1,100 extra requests round up to 2 blocks, adding $3.00 to the $9.99 subscription, before tax and provider charges. Its default safety limit is 22,000 requests.