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

# vLLM

> Configure the vLLM provider profile in Loong.

# vLLM

Local or self-hosted servers where the model boundary stays operator-owned. This page covers the built-in `vllm` provider kind directly.

## At A Glance

| Field            | Value                                       |
| ---------------- | ------------------------------------------- |
| Built-in kind    | `vllm`                                      |
| Provider group   | Local And Self-Hosted Providers             |
| Protocol family  | `openai_chat_completions`                   |
| Feature family   | `openai_compatible`                         |
| Auth scheme      | `bearer`                                    |
| Credential envs  | none required by default                    |
| Aliases          | `vllm_compatible`                           |
| Default base URL | `http://127.0.0.1:8000`                     |
| Request endpoint | `http://127.0.0.1:8000/v1/chat/completions` |
| Models endpoint  | `http://127.0.0.1:8000/v1/models`           |

## Minimal Config

```toml theme={null}
active_provider = "vllm"

[providers.vllm]
kind = "vllm"
base_url = "http://127.0.0.1:8000"
model = "auto"
```

## Verify It

```bash theme={null}
loong doctor
loong runtime models list
loong ask --message "Say hello and name the active provider."
```

If `runtime models list` is unreliable for this account or region, pin an explicit `provider.model` or add `preferred_models` instead of leaving recovery implicit.

## Auth And Routing Contract

| Contract                    | Value                                       |
| --------------------------- | ------------------------------------------- |
| Auth optional               | yes                                         |
| Model probe auth optional   | yes                                         |
| Default API key env         | none                                        |
| OAuth env                   | none                                        |
| Primary request route       | `http://127.0.0.1:8000/v1/chat/completions` |
| Primary model-catalog route | `http://127.0.0.1:8000/v1/models`           |

## Operator Notes

* Authentication is optional by default on this profile, which is useful for local or self-hosted servers during first setup.

## Related Docs

* Continue to [Providers And Models](/use-loong/providers-and-models) when you still need the broader chooser page.
* Continue to [Provider Guides](/use-loong/provider-guides/index) for the full built-in provider matrix.
* Continue to [Provider Recipes](/use-loong/provider-recipes) for representative rollout recipes.
* Continue to [Configuration Patterns](/use-loong/configuration-patterns) for the shared provider-profile shape.