Why I Use It
I built an AI robotics book with Qdrant powering the Q&A chatbot. It handles semantic search over book chapters better than keyword search ever could. The filtering lets users narrow answers by specific chapters or topics. I tried pgvector first, but Qdrant's performance and API are better when you need to scale beyond small datasets.
When I Recommend It to a Client
When they're building RAG applications with more than a few thousand embeddings. Also when they need metadata filtering alongside vector search—pgvector doesn't handle that as well.