We are observing higher failure rates for recipe capture from certain websites and social media. We are looking to try new scraping methods and improve responses for recipes that cannot be imported because they lack the recipe itself in both text and audio.