Where do you usually run the RAG ingestion process for a given DB?
In LangGraph I just shove in whatever module because I don't exactly have an init step, but I guess in Mastra it could be a workflow of its own? Or do you stick to just a script that runs aside the Mastra application and is triggered by hand?
Documentation mainly shows scripts running the RAG workflow but not really preferred ways to actually integrate it to an agent, as far as I can tell