Build a Production-Ready AI Transcription App

You are an AI/ML engineer and full-stack developer. Build a complete transcription service application powered by Whisper.

ARCHITECTURE:
- Tech Stack: Python FastAPI + React
- AI Integration: Whisper with proper API handling, error recovery, and fallback mechanisms
- Features: conversation history

REQUIREMENTS:
1. Complete application architecture with AI service layer
2. User interface for interacting with AI models
3. Prompt engineering and template system
4. Response streaming implementation
5. Conversation management and history
6. User authentication and usage tracking
7. Cost monitoring and budget alerts
8. Rate limiting and quota management
9. Error handling for API failures
10. Analytics dashboard for usage metrics
11. Admin panel for model configuration
12. API documentation

DELIVERABLES:
- Full-stack application code
- AI service integration layer
- Frontend with real-time updates
- Database schema for conversations and analytics
- API endpoints for AI interactions
- WebSocket setup for streaming (if applicable)
- Environment configuration
- Deployment guide
- Prompt library and examples

Generate a production-ready AI application with all necessary components, error handling, and best practices.

BONUS: Add troubleshooting section and common pitfalls to avoid.

NOTE: Focus on scalability, security, and best practices throughout.

SCALE: Design for handling millions of users/transactions. [Ref: 74d49ef9]