Problem
PubMed, JSTOR, government data portals all gate behind CAPTCHA. Long-running scrape agents lose hours when challenges fire.
Solution
Phone buzzes with the CAPTCHA visible. Researcher taps the I'm-not-a-robot box (or solves the tile puzzle). Agent resumes with cookies preserved.
Why it matters
30% of agent failures recovered without human attention loss
BatonPass does not solve CAPTCHAs autonomously. Every CAPTCHA solve is performed by a human operator, fully ToS-compliant.