What is an AI gateway?
An AI gateway is a single API that sits in front of multiple model providers. Instead of integrating each provider separately, you authenticate once, call models by name, and get consistent billing and usage visibility across all of them.
What does an AI gateway actually do?
It collapses several integrations into one. The work it removes is not the HTTP call — that is easy — but everything around it: separate credentials, separate billing relationships, separate usage reporting, and separate reconciliation at the end of the month.
- One credential instead of one per provider
- One request format, normally OpenAI-compatible
- One balance or invoice instead of several
- Usage and cost reported consistently across models
AI gateway vs LLM gateway
The terms are used interchangeably. "LLM gateway" tends to describe text models specifically, while "AI gateway" is sometimes used more broadly to include image, audio or embedding models. There is no formal distinction.
What to evaluate when choosing one
The differences that matter are structural rather than cosmetic, and most of them only become visible after you have traffic running.
- Deployment — hosted service, or a library you run yourself
- Billing — does it hold a balance, or do you keep provider accounts and pay each one
- Model selection — do you name the model, or does it choose for you
- Cost visibility — can you see the cost of one request, or only a monthly total
- Compatibility — how much of your existing client code survives
Where Astrodyne sits
Astrodyne is a hosted, OpenAI-compatible gateway with one prepaid balance and per-request cost visibility. You name the model and it calls that model — it does not substitute one for another — so behaviour and cost stay predictable.
Frequently asked questions
What is an AI gateway?
A single API in front of multiple model providers, handling authentication, request routing, billing and usage visibility so you do not integrate each provider separately.
What is an LLM gateway?
The same idea, usually said of text models specifically. The terms are used interchangeably and there is no formal distinction.
Do I still need provider accounts?
With a gateway that holds a balance, no — that is the point of the model. With a self-hosted proxy library you generally keep your own provider accounts and keys.
Does an AI gateway add latency?
A gateway adds one network hop between your application and the model provider. Whether that matters depends on the workload; for most applications it is small relative to the time the model spends generating a response.
Does Astrodyne choose a model for me?
No. You name the model in the request and Astrodyne calls that model. It never substitutes a different one, so cost and behaviour stay predictable.