When would you choose Karate's UI driver over a separate Selenium-based framework?

Karate's UI driver makes the most sense when a team already uses Karate for API testing and wants UI tests to share the same syntax, reporting, and CI integration rather than maintaining a second, unrelated framework and toolchain just for browser automation.

It's also a strong fit for workflows that genuinely span both layers, like logging in through the UI, then verifying the resulting state through a direct API call, since both parts of that test live in the same feature file using the same variable and assertion syntax, rather than coordinating between two separate frameworks that don't naturally share context.

A dedicated Selenium-based framework (or one of its modern wrappers) still makes sense when UI testing is the primary or sole focus of a team with deep existing Selenium expertise, extensive custom page-object infrastructure already built, or a need for very specific browser/driver combinations and plugins that a broader ecosystem's tooling supports more thoroughly than Karate's built-in driver does.