Consider this sort of query to AI: "Give me a list of SOMETHING, showing name and photo of each item."
Irrespective of AI model used, images either do not render (eg in case of ‘Kagi Quick’ model) OR do not exist (are omitted) (eg in case of ‘Gemini 3.1 Flash Lite preview’). An actual example (using the aforementioned models) is show below.
As you can see in the above example, Kagi Assistant does not render the images.
It would be really nice if Kagi Assistant could return images as a gallery (eg: CSS grid 🤔) with captions.
I dunno about video in responses, but it may be relevant to consider.
