Scrape image and video using AI prompt instead of using the cache

At the moment we have

[img-src-base64:%keyword%]

To generate an image based on a keyword

Need 2 new prompts to scrape an image or youtube video

[scraped-image:%keyword%]

[scraped-video:%keyword%]

This will allow us to insert images and videos related only to the sub heading of an article.

eg

<h2>%sub1%</h2>
[scraped-video:%sub1%]
[article about %sub1%]