Bing Behavior Automation with Fingerprints

I need a small-footprint automation script that opens Bing, performs realistic search sessions, and mimics genuine human behaviour at every step. The script must let me set a unique digital fingerprint for each visit (canvas, WebGL, timezone, language, etc.) and rotate user-agents freely so every request looks like it is coming from a different device and browser.

Primary goals
• Launch Bing searches, click a few organic results, scroll and dwell for randomised periods, then move on to the next query.
• Accept custom keyword lists fed in via a .txt or .csv file.
• Log each session’s full chain of actions (query, clicked URL, timestamps) so I can review how the browsing unfolded.

Technical notes
I am comfortable running headless or headed Chrome/Edge driven by Playwright or Selenium, but I am open to Puppeteer or any other stack you prefer as long as fingerprint injection and user-agent switching are rock-solid and detectable only as real traffic. Proxy support (HTTP/SOCKS, rotating) should be plug-and-play.

Acceptance criteria
• Unique fingerprint and user-agent confirmed through browserleaks.com (or similar) for each session.
• Search flow and click paths visible in the log file, matched to my supplied keyword list.
• Clean, well-commented source code plus a quick readme that lets me reproduce results on a fresh Windows or Linux box.

If you have past experience spoofing fingerprints or building search-engine behaviour bots, this should be straightforward. Let me know which language and libraries you prefer and any small demos you might have.