[pacarvalho]

Maybe not even computer vision on the pixel level but instead ML on the DOM to notice when it has loaded enough to parse the content from it?