OpenAI
OpenAI's GPT models and API services.
Configuration
PROVIDER=openai
API_KEY=sk-your-openai-api-key
MODEL=o3
TEMPERATURE=0.3
TOP_P=1.0
MAX_TOKENS=1024
REASONING_EFFORT=low
Key Parameters
| Parameter |
Description |
Default |
API_KEY |
OpenAI API key (required) |
- |
MODEL |
Model to use |
- |
TEMPERATURE |
Randomness (0.0-1.0) |
0.3 |
TOP_P |
Nucleus sampling |
1.0 |
MAX_TOKENS |
Max response tokens |
1024 |
TIMEOUT |
Request timeout (seconds) |
60 |
REASONING_EFFORT |
Thinking depth (0-10) |
- |
BASE_URL |
Custom API endpoint |
https://api.openai.com/v1 |
EXTRA_HEADERS |
Additional request headers |
- |
EXTRA_BODY |
Additional request parameters |
- |
Azure OpenAI Configuration
PROVIDER=openai_azure
API_KEY=your-azure-openai-key
ENDPOINT=https://your-resource.openai.azure.com/
API_VERSION=2024-02-01
DEPLOYMENT=your-deployment-name
MODEL=gpt-4
Azure Key Parameters
| Parameter |
Description |
Default |
API_KEY |
Azure OpenAI key |
- |
ENDPOINT |
Azure OpenAI endpoint |
- |
API_VERSION |
Azure OpenAI API version |
2024-02-01 |
DEPLOYMENT |
Azure deployment name |
- |
AD_TOKEN |
Azure AD OAuth token (optional) |
- |
DEFAULT_QUERY |
Default query parameters |
- |
Features
- ✅ Streaming responses
- ✅ Function calling
- ✅ MCP support
- ✅ Vision capabilities
- ✅ Reasoning parameters
- ✅ Azure integration
- ✅ High context windows
Important Notes
- API keys start with
sk-
- Rate limits apply based on your plan
- Azure requires additional configuration
- Models have different token limits and capabilities
- Pricing varies by model and usage
- Some models support image input
- Custom base URLs can be set for compatible endpoints
Resources