And when I get the answer I’d like to be

Hmm, lets back up a few steps haha

What kind of index are you using? How do you load the documents currently?

When you are loading data, you can set the extra_info field of each document to a dict of metadata

document.extra_info ={}

These get used in embeddings, and also show up in response source_nodes

response.source_nodes[0].node.node_info I think is the right way to access it lol