List of steps to reproduce (step by step, including full links if applicable):

• Attempt translation of a webpage in www.bloomberg.com

What happens?:
Selection steps relying on the page's HTML invariably return an unexpected output, because the server redirected us to an "Are you a robot?" webpage.

Citoid selection steps mostly work, but eventually they fail after repeated requests.

What should have happened instead?:
We should minimize redirections to this "Are you a robot?" page.

When I started testing, this was failing with XPath selection steps only. Citoid selections was working OK. I wondered whether including some of the headers included in the Citoid/Zotero request to their server (e.g., user-agent; see T302591) would help prevent/minimize this.

However, eventually, Citoid/Zotero requests would also start failing after repeated requests, both when running the Citoid service locally, or on Wikimedia's servers: T210871.

I wasn't able to hit the "Are you a robot?" page by repeatedly refreshing the page on my browser, though.