Hi everyone, I have a question about choosing a vector database for production use.
In what cases is Supabase with pgvector the better choice, and when is it better to use a dedicated vector database instead?
I’m trying to understand the trade-offs in areas like performance, scalability, filtering, operational complexity, cost, and developer experience. For example, if an app already uses Supabase/PostgreSQL, is pgvector usually enough, or are there clear situations where a dedicated vector DB becomes the better option?
I’d really appreciate practical advice based on real-world usage.
A user is seeking advice on choosing between Supabase with pgvector and a dedicated vector database for production use. They are interested in understanding the trade-offs in performance, scalability, filtering, operational complexity, cost, and developer experience. The user requests practical advice based on real-world usage scenarios.
https://supabase.com/blog/pgvector-vs-pinecone You might find this helpful which is a comparison between pg vector and one of the vector databases. Note that it is quite old and things may be very different now but still helpful. In practice all of the different vector databases have a pgvector vs ... article that you could probably read also