[spiralk]

I actually do something like this with my logseq notes. Since all of the files are .md in a directory, one can load them all in to a vectordb and use it for RAG. Logseq has an API too but using the .md files is easy.