[thenanyu]

You can prompt the model to dump all of the memory into a text file and import that.

In the onboarding flow, I can ask you, "Do you use another LLM?" If so, give it this prompt and then give me the memory file that outputs.