Both bugs share the same root cause: The code was written for happy-path, small-scale local testing — not for real users uploading large documents concurrently. The fixes are simple once you see…

Deleting for one user never breaks another user's search → Concurrent uploads are handled safely with no extra infra Vectors are shared. Access is isolated. That's the right model for multi-user RAG.…

The Roadmap for V2: Fixing the chunking problem. I will be building advanced strategies to parse non-linear PDFs: • Section-aware splitting (Methodology vs. Results) • Hierarchical node relationships…