I tried biloba on my existing playwright test suite and got a 2.5X improvement in runtime over playwright, and in the process improved a few flaky tests so they're now rock solid! And claude/codex do really really well working with Biloba, even more so than with playwright. Since I block my release pipeline on these tests, that speedup means I get to deploy changes much faster!