Icon for AI Gateway HQ

AI Gateway HQ

Free Trial

LLM gateway built around spend limits enforced before a request reaches a provider

AI Gateway HQ sits between applications, coding agents and automations and the model providers behind them, routing each request to an approved model and failing over when one is slow, unavailable or over budget.

Budgets are enforcement rather than reporting. Every eligible request reserves its ceiling before a provider can charge, and new provider attempts stop at the cap with automatic reload off until separately authorized.

With bring-your-own-key usage, customers pay their providers directly at their existing rates; AI Gateway HQ charges a separate gateway fee with no inference markup.

Hosting Cloud
Pricing Usage Based, Free tier, then $0.10/1,000 requests
HQ 🇺🇸 United States
Founded 2026
Compliance SSO
Screenshot of AI Gateway HQ webpage

What is AI Gateway HQ?

AI Gateway HQ is a hosted gateway that sits between an organization's AI tools and the model providers behind them. Applications, coding agents and automations send requests through it, and the gateway routes each one to an approved model, moving to a compatible backup when a provider is slow, unavailable or over budget. Providers covered include OpenAI, Anthropic, Bedrock and Gemini.

Budgets enforced before dispatch

The distinguishing feature is that budgets stop spend rather than report it. Every eligible request reserves its ceiling before an approved provider can charge, organization and workload limits sit next to their hard-stop behavior in the same console, and new provider attempts stop at the cap with automatic reload off until separately authorized.

Pricing

A no-card Test Lab runs policy logic without connecting a provider, with a $1 gateway-request credit that is not model credit.

Flex is prepaid with no monthly base, at $0.10 per 1,000 successful requests, with no charge for failed upstream calls. Company is $499 per month for 2,000,000 successful requests, adding OIDC, SAML, SCIM and built-in roles. Portfolio is $1,500 per month base for 10,000,000 requests, aimed at sponsors reporting across up to 100 approved companies. Hosted enterprise engagements start at $36,000 per year and private-deployment design at $60,000 per year.

The gateway fee does not scale with the model bill: it meters successful governed requests rather than taking a percentage of inference. Bring-your-own-key provider usage is billed by the provider directly at the customer's own rates, with no markup.

Who it fits

Teams where several people or tools use AI against one spending ceiling, and where switching models or provider accounts should not mean rewriting every client. The routing, workload keys and decision evidence are aimed at CIO, CISO and FinOps buyers rather than at individual developers.

Work on AI Gateway HQ? Feature it at the top of Inference APIs.

Is your product missing?

Add it here →