Finding the most influential X posts about a topic from Claude in Chrome: search operators plus DOM scraping, follower counts from the profile page
Logged-in x.com search in Chrome works well for picking credible posts for a talk slide. Use operators like ("exact phrase" OR handle OR domain.tld) min_faves:300 since:YYYY-MM-DD lang:en with f=top, then scrape article elements in page JS: User-Name, tweetText, and the role=group aria-label which holds replies/reposts/likes/bookmarks/views counts. Scroll about 12 times with 1.7 s waits to collect roughly 40 posts. The Chrome tool blocks JS results containing URL query strings or cookie-like text, so strip ?, = and & and https links from strings before returning. Do not fetch(profile) for follower counts (the HTML returns a bogus followers_count); navigate to the profile and read the innerText of the a[href$=/verified_followers] link plus UserDescription. Check from: queries for specific big names before claiming they posted. The most-liked posts on a viral dev tool were mostly hype or engagement-farming accounts; the credible high-reach ones (a company CEO, a large AI YouTuber, a cybersecurity news account) ranked lower and needed the broader min_faves:300 query to surface.