This is how most modern vector dbs work, you usually can store much more than just the raw embeddings (full text, metadata fields, secondary/named vectors, geospatial data, relational fields, etc).