The OpenAI Python SDK requires authentication using an API key. This guide covers how to set up authentication for both OpenAI and Azure OpenAI.
OpenAI API Key
Environment Variable (Recommended)
The most secure way to provide your API key is through the OPENAI_API_KEY environment variable:
export OPENAI_API_KEY="sk-proj-..."
from openai import OpenAI
client = OpenAI()
# API key is read from OPENAI_API_KEY environment variable
Using python-dotenv
For local development, use python-dotenv to load your API key from a .env file:
Install python-dotenv
pip install python-dotenv
Create .env file
Create a .env file in your project root:OPENAI_API_KEY="sk-proj-..."
Add .env to your .gitignore to avoid committing your API key to version control.
Load environment variables
import os
from dotenv import load_dotenv
from openai import OpenAI
load_dotenv() # Load variables from .env
client = OpenAI()
# API key is now loaded from .env file
Direct API Key (Not Recommended)
You can pass the API key directly, but this is not recommended for production:
from openai import OpenAI
client = OpenAI(
api_key="sk-proj-...", # Not recommended
)
Never hardcode API keys in your source code or commit them to version control.
Get an API Key
You can obtain an API key from the OpenAI Platform.
Azure OpenAI Authentication
Azure OpenAI supports two authentication methods: API keys and Azure Active Directory (Azure AD).
API Key Authentication
Use the AZURE_OPENAI_API_KEY environment variable:
export AZURE_OPENAI_API_KEY="your-azure-key"
export AZURE_OPENAI_ENDPOINT="https://example-endpoint.openai.azure.com"
export OPENAI_API_VERSION="2023-07-01-preview"
from openai import AzureOpenAI
client = AzureOpenAI(
api_version="2023-07-01-preview",
azure_endpoint="https://example-endpoint.openai.azure.com",
)
Azure Active Directory (Azure AD)
For enhanced security, use Azure AD tokens:
from azure.identity import DefaultAzureCredential, get_bearer_token_provider
from openai import AzureOpenAI
token_provider = get_bearer_token_provider(
DefaultAzureCredential(),
"https://cognitiveservices.azure.com/.default"
)
client = AzureOpenAI(
api_version="2023-07-01-preview",
azure_endpoint="https://example-endpoint.openai.azure.com",
azure_ad_token_provider=token_provider,
)
Organization and Project IDs
You can optionally specify organization and project IDs:
from openai import OpenAI
client = OpenAI(
organization="org-...", # Optional
project="proj_...", # Optional
)
export OPENAI_ORG_ID="org-..."
export OPENAI_PROJECT_ID="proj_..."
Callable API Keys
For advanced use cases where you need to refresh tokens, you can pass a callable that returns the API key:
from openai import OpenAI
def get_api_key() -> str:
# Fetch or refresh your API key
return "sk-proj-..."
client = OpenAI(
api_key=get_api_key,
)
