My existing AI-driven web search agent was originally wired to “LinkUp” for crawling, but that service no longer meets my needs. I now need the entire search layer refactored to run on Brightdata so I can continue gathering research-grade data with real-time updates. Here’s what has to happen: • Swap every LinkUp dependency for Brightdata, including authentication, proxy rotation and rate-limit handling. • Preserve (or improve) the current accuracy logic so result quality does not slip. • Activate real-time data updates—the agent should push fresh results automatically without manual triggering. • Provide a concise setup guide and any modified source files so I can redeploy quickly. A successful hand-off is one where the agent returns the same or better volume of research data, updates continuously, and runs without breaking any existing routines.