Congratulations on launch! I've seen your tweet about unlimited context for LLMs :)

Do you have option to manually provide text that will be used as LLM memory?