Changelog:
- Refactored LLM service for better error handling and content parsing
- Updated backend API endpoint to increase overall concurrent request limit