- Refactored AI gateway to utilize new cost management structures for usage tracking.
- Replaced deprecated token extraction methods with a unified cost parsing approach.
- Enhanced usage fallback mechanisms and introduced detailed usage metrics in responses.
- Added new metering functionality to record AI usage and costs effectively.
- Updated tests to reflect changes in usage parsing and cost calculations.
- Introduced new API endpoints for retrieving AI usage summaries and pricing information.
- Updated .env.example to include new configuration options for the UltiAI branding and API endpoints.
- Enhanced Nginx configuration to support new API routes for the MCP and WebSocket connections.
- Introduced sub-filters for branding adjustments in Nginx responses.
- Added new JavaScript patch for API endpoint adjustments.
- Implemented tests for new API functionalities and improved error handling in the AI gateway.
- Updated .env.example to include new configuration options for AI gateway and WebUI secret key.
- Modified Nginx configuration to support additional API routes for model management and migration.
- Implemented new API endpoints for discovering organization-level LLM models and managing hosted mail services.
- Enhanced AI gateway logic to support organization-specific model access and permissions.
- Improved error handling and response structures in the AI and mail APIs.
- Added integration tests for new features and updated existing tests for model access control.
