Creor/Docs
Dashboard

Azure OpenAI

Azure OpenAI lets you run GPT and o-series models on Azure infrastructure with enterprise-grade security, compliance, and regional data residency.

Available Models

Azure OpenAI hosts the same models as OpenAI but through Azure's infrastructure. Available models depend on your Azure region and deployment.

ModelDeployment Name (typical)Best For
GPT-4ogpt-4oGeneral-purpose coding, fast responses
GPT-4o Minigpt-4o-miniCost-effective lightweight tasks
GPT-4 Turbogpt-4-turboComplex coding with vision support
o3o3Advanced reasoning, complex architectures
o3-minio3-miniFast reasoning at lower cost
o4-minio4-miniNewest reasoning model

Note

In Azure OpenAI, you create "deployments" with custom names. The model ID you use in Creor is your deployment name, not the base model name. For example, if you deployed GPT-4o as "my-gpt4o", you would use that as the model ID.

Setup

Prerequisites

  • An Azure subscription with Azure OpenAI access approved.
  • An Azure OpenAI resource created in the Azure portal.
  • At least one model deployed in your Azure OpenAI resource.

Step 1: Get Your Credentials

From the Azure portal, navigate to your Azure OpenAI resource and collect:

  • Endpoint URL -- found in the "Keys and Endpoint" section (e.g., https://your-resource.openai.azure.com).
  • API Key -- either Key 1 or Key 2 from the same section.
  • Deployment name -- the name you gave your model deployment.

Step 2: Configure Environment Variables

1
2
3
export AZURE_OPENAI_API_KEY="your-azure-api-key"
export AZURE_OPENAI_ENDPOINT="https://your-resource.openai.azure.com"
export AZURE_OPENAI_DEPLOYMENT="your-deployment-name"

Step 3: Set the Model in Creor

Reference Azure models using the azure/ provider prefix followed by your deployment name:

1
2
3
{
"model": "azure/your-deployment-name"
}

Configuration

Basic Configuration

1
2
3
4
5
6
7
8
9
10
{
"model": "azure/gpt-4o",
"provider": {
"azure": {
"options": {
"baseURL": "https://your-resource.openai.azure.com/openai"
}
}
}
}

Multiple Deployments

Use different Azure deployments for different agents:

1
2
3
4
5
6
7
8
9
10
{
"agent": {
"build": {
"model": "azure/gpt-4o-deploy"
},
"plan": {
"model": "azure/o3-deploy"
}
}
}

Completion URLs Mode

If your Azure deployment uses the older completions API instead of the responses API, enable the useCompletionUrls option:

1
2
3
4
5
6
7
8
9
{
"provider": {
"azure": {
"options": {
"useCompletionUrls": true
}
}
}
}

Azure Cognitive Services

Creor also supports the Azure Cognitive Services endpoint variant, which uses a different base URL pattern. This is common in enterprise Azure setups.

1
2
3
4
5
# Set the resource name
export AZURE_COGNITIVE_SERVICES_RESOURCE_NAME="your-resource-name"
 
# The endpoint becomes:
# https://your-resource-name.cognitiveservices.azure.com/openai

Use the azure-cognitive-services provider prefix:

1
2
3
{
"model": "azure-cognitive-services/your-deployment-name"
}

Advanced Options

API Version

Azure OpenAI APIs are versioned. If you need a specific API version:

export AZURE_OPENAI_API_VERSION="2024-12-01-preview"

Request Timeout

1
2
3
4
5
6
7
{
"provider": {
"azure": {
"timeout": 600000
}
}
}

Model Filtering

1
2
3
4
5
6
7
{
"provider": {
"azure": {
"whitelist": ["gpt-4o-deploy", "o3-deploy"]
}
}
}

Best For

  • Organizations that must keep traffic within Azure infrastructure.
  • Teams using Azure Active Directory for access control.
  • Deployments requiring regional data residency (EU, Asia, etc.).
  • Enterprise compliance requirements (SOC 2, ISO 27001, HIPAA).
  • Companies with existing Azure OpenAI commitments or reserved capacity.

OpenAI (Direct)

Use OpenAI models directly through the OpenAI API.

AWS Bedrock

Run Claude and other models on AWS infrastructure.

PreviousAWS BedrockModels & ProvidersNextOpen Source & OthersModels & Providers