# What is an LLM gateway?

An LLM gateway is one endpoint in front of every model: your code calls a single OpenAI-compatible API, and the gateway handles authentication, routing, metering, failover, and policy across providers. It turns model choice from a code dependency into configuration.

The gateway is where production concerns live: per-key spend caps, per-route metering, provider failover, and audit. Application code stays a stable API call while models, providers, and prices change behind it.

Overhead is the design constraint that separates gateways: the hop must cost milliseconds, not another model’s worth of latency. Allocate’s gateway holds itself to a 30 ms median overhead budget, reported on every response.

## See also

- [Browse the model catalog](https://allocate.network/models)

## Related terms

- [Inference](https://allocate.network/glossary/inference.md)
- [Model routing](https://allocate.network/glossary/model-routing.md)
- [OpenAI-compatible API](https://allocate.network/glossary/openai-compatible-api.md)

---

[HTML page](https://allocate.network/glossary/llm-gateway) · [Machine-readable catalog](https://allocate.network/catalog.json)