It seems I have to use page.should have_content("...") instead of response.should have_content("...") Maybe that's because Capybara is the default engine now and not Webrat? -- Posted via http://www.ruby-forum.com/.