"author:nathanj2570" matching MCP servers:

Enables AI agents and humans to share one live, URL-synced mobile and desktop browsing session, where the agent can navigate, tap, type, scroll, switch devices and read notes the human pins on the screen. Instead of guessing from screenshots, it computes real measurements from the live DOM — contrast ratios, undersized touch targets, clipped or covered controls, overflow and dark-mode drift — and returns them as text findings with the offending selector attached.