Fix HuggingFace Spaces deployment and add GPT-5 support

Implement proper async parallelism for multi-agent pipeline

Fix HuggingFace Spaces import error with lazy loading

Update README and enhance query processing with logging and error handling

Refactor Gradio UI components and add configuration class for improved maintainability

Disable HF token for CrossEncoder to avoid auth errors with invalid tokens

Fix CrossEncoder API - use 'token' parameter instead of deprecated 'use_auth_token'

Add HuggingFace token support for cross-encoder model loading

Force HuggingFace to pull latest __init__.py with optional OpenSearch imports

Fix imports to avoid OpenSearch dependencies in memory-only mode

Update requirements and enhance MemoryRetriever with cross-encoder reranking; improve retrieval logic and logging

Refactor document ingestion and processing; update configurations for chunking and retrieval, enhance error logging, and implement markdown-aware chunking

Update dependencies and refactor agent initialization; add optional fact-check skipping for performance improvements

Refactor RAG Email Assistant for in-memory processing; update configurations, implement memory indexing and retrieval, enhance Gradio UI, and streamline document ingestion.

Implement modular RAG email assistant architecture