---
title: "Providers"
subtitle: "Every provider the gateway calls, and how to pick one."
---

## Change the provider

A model name such as `gpt-5.5` can be served by any provider that has it. Put a provider's prefix in front, such as `azure/gpt-5.5`, to keep the call on that provider. To pick among providers by price, speed, or data retention instead, see [Routing](/docs/ai-gateway/routing/provider-routing).

<Tabs>
<Tab title="cURL">
```bash
curl https://gateway.respan.ai/v1/chat/completions \
  -H "Authorization: Bearer $RESPAN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "azure/gpt-5.5", "messages": [{"role": "user", "content": "Hello!"}]}'
```
</Tab>
<Tab title="Python">
```python
import os
from openai import OpenAI

client = OpenAI(base_url="https://gateway.respan.ai", api_key=os.environ["RESPAN_API_KEY"])

# Any provider that serves gpt-5.5
client.chat.completions.create(model="gpt-5.5", messages=[{"role": "user", "content": "Hello!"}])
# Azure OpenAI only
client.chat.completions.create(model="azure/gpt-5.5", messages=[{"role": "user", "content": "Hello!"}])
```
</Tab>
<Tab title="TypeScript">
```typescript
import OpenAI from "openai";

const client = new OpenAI({ baseURL: "https://gateway.respan.ai", apiKey: process.env.RESPAN_API_KEY });

// Any provider that serves gpt-5.5
await client.chat.completions.create({ model: "gpt-5.5", messages: [{ role: "user", content: "Hello!" }] });
// Azure OpenAI only
await client.chat.completions.create({ model: "azure/gpt-5.5", messages: [{ role: "user", content: "Hello!" }] });
```
</Tab>
</Tabs>

## Provider list

Put a provider's pin prefix in front of a model ID to keep the call on that provider. To run on your own account with a provider, add its key under **API keys** › **BYOK** (Team and above); your own keys serve first.

| Provider | Pin prefix | Your own key needs |
|---|---|---|
| OpenAI | `openai/` | API key |
| Anthropic | `anthropic/` | API key |
| Google Gemini | `gemini/` | API key |
| xAI | `xai/` | API key |
| AWS Bedrock | `bedrock/` | AWS access key, secret key, region |
| Google Vertex AI | `vertex_ai/` | Project, location, service account JSON |
| Azure OpenAI | `azure/` | API key, endpoint URL, API version |
| Mistral | `mistral/` | API key |
| DeepSeek | `deepseek/` | API key |
| Groq | `groq/` | API key |
| Cohere | `cohere/` | API key |
| Perplexity | `perplexity/` | API key |
| Together AI | `togetherai/` | API key |
| Fireworks AI | `fireworks/` | API key |
| Cerebras | `cerebras/` | API key |
| OpenRouter | `openrouter/` | API key |
| Replicate | `replicate/` | API key |
| Baseten | `baseten/` | API key, endpoint URL |
| Moonshot | `moonshot/` | API key |
| Nebius | `nebius/` | API key |
| Novita AI | `novita/` | API key |
| FriendliAI | `friendliai/` | Endpoint URL, API key |
| Parasail | `parasail/` | Endpoint URL, API key |
| Tera | `tera/` | Endpoint URL, API key |
| Inference.net | `inference_net/` | API key, endpoint URL |
| AI21 Labs | `ai21_labs/` | API key |
| Wafer | `wafer/` | Endpoint URL, API key |
| Nextbit | `nextbit256/` | API key, endpoint URL |
| TypeSafe | `typesafe/` | API key |
| Azure DeepSeek | `azure_deepseek/` | API key, endpoint URL |
| [Custom](#custom-providers) | The provider ID you set | Endpoint URL, API key |

## Custom providers

To call a self-hosted or other OpenAI-compatible endpoint, go to **API keys** › **BYOK**, click **New custom provider**, and enter a **Provider ID**, a **Provider name**, and the endpoint's **Base URL**, which must be a public address. Add its key with **Add deployment**. Then, for each model it serves, go to **Gateway** › **Models**, click **New custom model**, pick your provider, and send the **Model ID** the dialog shows as `model`.

## Become a Respan provider

To offer your models to everyone on Respan, reach out to [team@respan.ai](mailto:team@respan.ai).
