Comment by m3kw9

How does this work because the browser needs to render a page and the vision model needs to know where a button is, so it still needs to see an image. How does headless make it easier?