In my experience scraping both, APIs have tended to be more stable than the structure of HTML pages. I'd guess this is because there's an incentive to keep your

In my experience scraping both, APIs have tended to be more stable than the structure of HTML pages. I'd guess this is because there's an incentive to keep your API stable (a breaking change for my scraper is also a breaking change for the front-end), whereas there's basically no reason not to rewrite giant chunks of the HTML to accommodate a design change.