# Models endpoint

Call public `GET /v1/models` at runtime and select an identifier from the returned list instead of assuming a fixed catalog.

```bash
curl https://api.foxora.ai/v1/models
```

The response follows a list shape with model entries under `data`. Each entry includes an `id` suitable for supported requests and may include ownership or description fields.

1. Refresh discovery when the application starts or after an availability error.
2. Present returned descriptions without converting them into guaranteed provider names.
3. Send the selected `id` as the chat request’s `model`.
4. Handle a model becoming unavailable between discovery and execution.

The endpoint is public, but using a returned model for generation still requires authentication and the account’s applicable plan entitlement.

**Expected result:** discovery returns a non-empty list, and a permitted returned identifier can be used in `/v1/chat/completions`.