Clean up startup diagnostics and debug route

Simplify debug_cv2() response

Add debug endpoint for cv2 diagnostics

Add cv2 startup diagnostics

Add trace flushing for Langfuse telemetry

Switch LLM model and include usage in responses

Fix Langfuse, add LOGFIRE env

guardrails were not working. corrected that

Avoid RAG searches for greetings in system prompt

Update llm.py

Switch to local free HuggingFace embeddings

Switch decommissioned model to active llama-3.1-8b-instant

used groq api insteal of open router api

Restore: Switch back to openai/gpt-4o-mini

Update chat.py

Fix: Strip literal quotes from environment variables in tracing

Fix: Switch to free Gemini model for OpenRouter

done

Add session persistence and login_method metadata

faceid

ui

Improve auth and face service error handling

Scope RAG search to user and improve uploads

Restructure backend modules and update imports

Use shared DB, improve RAG tools and UI

Refactor backend: simplify logic & cleanup

Add session-based chat history support

frontend integration

Add backend features and exclude local frontend folder

Add Voice RAG AI backend and Streamlit UI