So you passed in a URL (that loads fine in your browser) as the 'contentUri' to your Rosette call -- and got a 400. What happened is that our underlying components in the Rosette use a custom URL text extractor, and sometimes sophisticated sites block our extractor, thinking it's a bot. We're working on catching these cases, and upcoming releases will have improved URL downloading. For now, the workaround is to take the text yourself from the webpage and pass it as text in the 'content' parameter.