AI Gateway HQ
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.
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.
AI Gateway HQ Alternatives
Explore 104 products in the Inference APIs category. View all AI Gateway HQ alternatives.
2kw.ai
EU-hosted AI integration platform combining a provider gateway, document extraction, evals and cost observability
Varion
OpenAI-compatible proxy that trims LLM token usage before requests reach your provider
AISIX
Open-source AI gateway in Rust with one OpenAI-compatible API across 100+ LLM providers
Parity Layer
Proves a cheaper model matches your current one on your own prompts, then routes to it
Also listed in
Work on AI Gateway HQ? Feature it at the top of Inference APIs.
Is your product missing?