Release v0.190.3
Key Updates & Enhancements
🔹 Role-Based Access Control (RBAC) Enhancements
- Admin and User Role Assignments:
- Introduced role-based access in the registered app.
- Users do not have access to Admin settings, ensuring better security and separation of privileges.
- New App Roles Setup Required:
- Users
- Allowed Member Types: Users/Groups
- Value:
User - Description: Normal user
- Admins
- Allowed Member Types: Users/Groups
- Value:
Admin - Description: Manage the application
- Users
- Assigning Roles:
- Navigate to Azure Portal → Enterprise Applications → Select the app associated with your registration.
- Go to Users and Groups → Assign users to their respective roles (
UserorAdmin). - This replaces the default access mechanism, ensuring proper permission management.
🔹 Simplified API Configuration
- GPT, Embedding, and Image Generation Endpoints:
- Configuration streamlined for easier setup and management.
- Reduces complexity in defining model endpoints while maintaining flexibility.
🔹 Improved OpenAI Model Access
- Registered App Permissions:
- The registered app now requires Cognitive Services OpenAI User permissions.
- This allows the application to fetch the available models dynamically from the configured endpoints.
🔹 AI Model Tracking for Responses
- Tracked Model Metadata:
- Every AI-generated response now includes metadata specifying the model used.
- Helps with debugging, auditing, and optimizing model performance.
Deployment Notes
- Ensure the App Roles (Users & Admins) are properly configured before deployment.
- Assign users to their respective roles through the Enterprise Application settings in Azure AD.
- Update configurations to reflect the simplified endpoint structure.
- Validate that role-based permissions are correctly applied in the admin settings.
This release focuses on security, structured role management, configuration simplification, and AI model tracking, making the platform more robust and scalable. 🚀