Azure OpenAI (gateway)

1
{
2
  "mcpServers": {
3
    "respan-docs": {
4
      "url": "https://mcp.respan.ai/mcp/docs"
5
    }
6
  }
7
}

Use Respan Gateway to call Azure OpenAI models while keeping unified observability (logs, cost, latency, and reliability metrics) in Respan.

Prerequisites

• A Respan API key
• Azure OpenAI credentials (BYOK)

Supported SDKs / integrations

Configuration

There are 2 ways to add your Azure OpenAI credentials to your requests:

Via UI (Global)

1

Navigate to Providers

Go to the Providers page. This page allows you to manage credentials for supported model providers.

2

Add your Azure OpenAI credentials

Select Azure OpenAI and add the required credential fields.

• api_key — your Azure OpenAI API key
• api_base — your Azure OpenAI resource endpoint, also called base URL
• api_version — the Azure OpenAI API version to use

3

Find credentials in Azure Portal

Visit the Azure Portal OpenAI Service, choose your deployment, and open Endpoints and Manage Keys to copy the endpoint and API key.

4

Map your deployment name

Azure allows custom deployment names. Use the azure/{deployment-name} model ID format in gateway requests.

5

Choose available models (Optional)

Copy the model ID from the Respan Models page, paste it into the available models field, and press Enter. Leave the field empty to apply the credentials to all Azure OpenAI models.

6

Configure Load Balancing (Optional)

Add multiple Azure OpenAI credential sets for redundancy. Use the Load balancing weight field to control traffic distribution.

Via code (Per-Request)

You can pass credentials dynamically in the request body. This is useful if you need to use your users’ own API keys (BYOK).

Add the customer_credentials parameter to your Gateway request:

1
{
2
  "customer_credentials": {
3
    "azure_openai": {
4
      "api_key": "YOUR_AZURE_OPENAI_API_KEY",
5
      "api_base": "https://your-resource.openai.azure.com/",
6
      "api_version": "2024-10-21"
7
    }
8
  }
9
}

Override credentials for a particular model (Optional)

Use credential_override when one request or model should use different credentials than the default provider key.

1
{
2
  "customer_credentials": {
3
    "azure_openai": {
4
      "api_key": "YOUR_AZURE_OPENAI_API_KEY",
5
      "api_base": "https://your-resource.openai.azure.com/",
6
      "api_version": "2024-10-21"
7
    }
8
  },
9
  "credential_override": {
10
    "azure/gpt-4o": {
11
      "api_key": "ANOTHER_API_KEY",
12
      "api_base": "https://another-resource.openai.azure.com/",
13
      "api_version": "2024-10-21"
14
    }
15
  }
16
}

Supported models

Find the complete and current list of Azure OpenAI model IDs on the Respan Models page. Use the exact model ID shown there in your gateway requests.

Log Azure OpenAI requests

If you are not using the Gateway to proxy requests, you can still log your Azure OpenAI requests to Respan asynchronously. This lets you track cost, latency, and performance metrics for external calls.

1
import requests
2
3
url = "https://api.respan.ai/api/request-logs/create/"
4
payload = {
5
  "model": "azure/gpt-4o",
6
  "prompt_messages": [
7
    {
8
      "role": "user",
9
      "content": "Hello, how are you?"
10
    }
11
  ],
12
  "completion_message": {
13
    "role": "assistant",
14
    "content": "Hello from Azure OpenAI through Respan."
15
  },
16
  "cost": 0.001,
17
  "generation_time": 1.2,
18
  "customer_params": {
19
    "customer_identifier": "user_123"
20
  }
21
}
22
headers = {
23
    "Authorization": "Bearer YOUR_RESPAN_API_KEY",
24
    "Content-Type": "application/json",
25
}
26
27
response = requests.post(url, headers=headers, json=payload)