###### Set up Respan

1. **Sign up** — Create an account at [platform.respan.ai](https://platform.respan.ai/)
2. **Create an API key** — Generate one on the [API keys page](https://platform.respan.ai/platform/api/api-keys)
3. **Add credits or a provider key** — Add credits on the [Credits page](https://platform.respan.ai/platform/api/billing) or connect your own provider key on the [Integrations page](https://platform.respan.ai/platform/api/providers)

###### Use AI

Add the [Docs MCP](/content/docs/documentation/get-started/mcp/docs-mcp/index.html) to your AI coding tool to get help building with Respan. No API key needed.

```

```

Respan supports models across major providers. This page covers how to choose between them per request, override provider credentials, and create custom model aliases. Use the [model catalog](/content/docs/documentation/features/gateway/models-catalog/index.html) for the current inventory and billing mode.

* * *

## Inline router (`models`)

Pass an inline list of candidate models per request and let the LLM router pick one. This is the request-time alternative to a pre-configured `load_balance_group` (see [Load balancing](/content/docs/documentation/features/gateway/load-balancing/index.html)).

```

```

| Field                | Type   | Description                                                   |
|----------------------|--------|---------------------------------------------------------------|
| `models`            | array  | Candidate model names. Router selects one per request.       |
| `exclude_models`    | array  | Models to exclude from selection.                             |
| `exclude_providers`  | array  | Providers to exclude from selection.                          |

The selected model is recorded on the log under `model`.

* * *

## Per-model credential override (`credential_override`)

Override credentials for a specific model on a single request. More granular than `customer_credentials` (which applies per provider). Useful when one model in a fallback chain needs different credentials than the rest.

```

```

The key is the full model slug (`azure/gpt-4o`, `vertex_ai/claude-sonnet-4-5@20250929`, etc.). Each fallback attempt resolves credentials per-model.

* * *

## Pin a provider

Use the `X-Respan-Route-Provider` header to route a request to a specific provider without changing the model slug. Useful when a CLI agent or SDK sends a fixed slug (e.g. `claude-sonnet-4-5-20250929`) and you want to route it to Vertex AI, Bedrock, or another provider.

```

```

* * *

## Custom model aliases

Sometimes you may want to:

- Use your own naming conventions (aliases) for existing models.
- Add a newly released model from a supported provider before we’ve listed it.

Create a **custom model** in your model list to do either.

### Why use a custom model?

Custom models let you create multiple aliases for the same underlying model.

This is useful when:

- Different use cases: Reuse the same base model but track usage separately.
- Analytics: Monitor usage by alias, not just the raw model name.
- Try new models before we’ve listed them.

### How to create a custom model

1. Go to the Models page and select `+ Create custom model`.
2. Fill out the form: model name, provider, costs, modality, etc.
3. Save. Your custom model will now appear in the model list.
4. Copy the model ID and use it in your code.

### Copy from a base model

Instead of starting from scratch, you can also copy a base model and give it a new name. This saves time if you just want multiple aliases of the same model without reconfiguring everything.

### FAQ

**Can I use a custom model in the LLM Gateway?** Yes, as long as it maps to a valid provider model. Example: if OpenAI launches `gpt-6`, you can add it as a custom model and immediately call it through the AI Gateway.

**When should I use “copy base model”?** Think of it as a preset. It’s the fastest way to create aliases without re-entering configuration details.