A useful way to think about this: browser-use is the runner, Sentience is the assertion layer. Agents act → Sentience verifies. If you’ve built flaky E2E test

A useful way to think about this: browser-use is the runner, Sentience is the assertion layer. Agents act → Sentience verifies. If you’ve built flaky E2E tests or agent demos that “usually work”, this is an attempt to make those workflows inspectable and testable.