Would it be possible to add an
is_quick_response
flag to the List Chat Events API so we can clearly distinguish between events generated by the Quick Response model versus the supplemental LLM?
Since the Quick Response model doesn't consume the system prompt and behaves differently from the supplemental LLM, having this distinction in the chat events would really help us. Right now it's difficult to infer this reliably from the event data alone.
Would love to know if this is something you'd consider adding, or if there's an existing way to differentiate these events that we may be missing.