Text with some images -> Some client writes me a message "my bike is broken! pedals make noise!" and I want my app to generate answer based on a bunch of manuals (including text with a lot of photos) "Dear Peter, you need to grease the pedal bearing, as shown on this photo [photo1.jpg]" - is there a solution for this? In other words, can LLM generate text+image answers, with images put in between the text paragraphs, just like in a typical manual?
Hmm, I feel like if you included a related image in the metadata of nodes, this could be possible? So that if that node is used to answer a query, show the user the related images?
