Python RAG/LLM application experiments with crawler-to-context flow, model calls, object storage configuration, and FastAPI endpoints.
Do not commit API tokens, crawler credentials, object storage keys, vector DB credentials, private source data, generated answers from protected content, or proprietary logs. Prefer sanitized fixtures and public reproduction notes.