You can read the original post over on lukaswalter.dev .

Developers often start RAG apps using tutorials that recommend dedicated vector databases.

Step 1: Sign up for a vector database like Pinecone or Qdrant.

This adds a costly SaaS service to your architecture or requires you to manage it yourself.

And if you are building line-of-business applications in .NET, dedicated vector databases often introduce another problem: Data Synchronization.

If core entities like Products, Customers, or SupportTickets exist in a relational database and vector embeddings reside in a specialized vector DB, you face a distributed systems challenge. What if a product is deleted or its description updated? Synchronizing datastores becomes daunting.

A pragmatic solution? Store your vectors alongside your relational data.

Using PostgreSQL, the pgvector extension transforms your relational database into a powerful vector search engine.…