[feat] `vector_strategies` for vector db indexing

It would be really cool to add indexing strategies for this data later on. Right now we have `storage_strategies`, which is fine, but I'm thinking of something like `vector_strategies` with `pineconedb`, `pg_vector`, and similar approaches—so that PDF documents are stored directly in a vector database. How awesome would that be? Then you could immediately set up a RAG (Retrieval-Augmented Generation) workflow!

Start by locating the existing `storage_strategies` implementation and its tests or entry points. From there, determine how indexing strategies are configured and what integration points would be needed for the proposed vector-database approaches. Done would require an agreed scope and implementation plan for supported vector stores and RAG-oriented document storage.

Written by the indexing model from the issue text.