- Introduced a timeout constant for LLM requests to ensure stability.
- Updated LLMFactory to set a default timeout if not provided.
- Enhanced streaming events to track started nodes for better UI feedback.
- Refactored lite discovery and enrichment nodes to utilize new search tool functions, improving query handling and result fetching.
- Added caching mechanisms for trusted source lookups to optimize performance.
- Adjusted various timeout settings across HTTP requests for consistency.