Very useful👍 thanks for sharing! QQ: from the logic you shared, isn't the chatbot only answering based on the user's comment and not the original post message or image/video? Is there a way to include all of this in the input for better context for chatGPT and more appropriate responses?
Yes, the flow would be a bit more complex. But a simple mapping function that finds the post based on where the comment was located, analyzes the post / image, and stores that info in the response prompt. Might need to be done internally with a db or post mapping but should be able to.
Very useful👍 thanks for sharing! QQ: from the logic you shared, isn't the chatbot only answering based on the user's comment and not the original post message or image/video? Is there a way to include all of this in the input for better context for chatGPT and more appropriate responses?
Yes, the flow would be a bit more complex. But a simple mapping function that finds the post based on where the comment was located, analyzes the post / image, and stores that info in the response prompt. Might need to be done internally with a db or post mapping but should be able to.
Hlo
👋