Creating RAG based agent reporting on my email data

I saw the list of AI enabled features, like them all, however I have yet another question: I would like to create a RAG agent, which would be a standalone application (not part of the emclient app). In order to do this, I would have to "ingest " all of emclient app data and create a “vector database” (corpus) that would be queried by this RAG agent.
Such design would allow me to later add more vector databases (corpora) and increase the agent’s usefulness.

Sub Questions:

  1. Do I ingest data from local data - where I am using four different mailboxes?
  2. How will I know that the local data is complete (synced with emclient’s database)?
  3. Has someone else already made a similar solution?