> ## Documentation Index
> Fetch the complete documentation index at: https://docs.sailresearch.com/llms.txt
> Use this file to discover all available pages before exploring further.

# US-Only Inference

> Restrict inference requests to United States capacity

<Note>
  US-only inference is available on the [Pro and Enterprise
  plans](/pricing#plans).
</Note>

Add `routing.allowed_countries: ["US"]` to run an inference request on United
States compute.

```bash theme={null}
curl https://api.sailresearch.com/v1/responses \
  -H "Authorization: Bearer $SAIL_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "zai-org/GLM-5.3-Flash",
    "input": "Explain the key ideas behind transformers.",
    "metadata": {
      "completion_window": "asap"
    },
    "routing": {
      "allowed_countries": ["US"]
    }
  }'
```

The restriction applies per request. Omitting `routing.allowed_countries` or
passing an empty list leaves the request unrestricted by country. Only `US` is
supported. Unsupported countries or malformed values return HTTP `400` with
error type `invalid_request_error`.

## Pricing

The surcharge is added to the selected model and completion window's
[per-token rates](/pricing#inference).

| Plan       | US-only surcharge |
| ---------- | ----------------: |
| Pro        |               20% |
| Enterprise |               10% |

See the [plan comparison](/pricing#plans) for the other benefits included with
each plan.

## When US capacity is unavailable

US-only requests never fall back outside the United States.

* `asap` returns HTTP `429`. Retry after the delay in the `Retry-After` header.
* `balanced` and `flex` wait for US capacity until the request deadline. Set
  `background: true` to submit and poll for the result.

## Model availability

Use a model from Sail's [model catalog](/models) and a [completion window](/completion-windows)
it supports. If your workload has specific capacity needs, [contact support](https://www.sailresearch.com/support).