While platforms such as Amazon Mechanical Turk have generated excitement as a potential source of income in developing regions, to date there remains little evidence that such opportunities have transformed livelihoods for low-income workers. In this study, we analyze the usability barriers that prevent those with basic digital literacy skills from accomplishing simple tasks on Mechanical Turk. Based on our observations, we design new user interfaces that reduce the barriers to task comprehension and execution. Via a study of 49 low-income workers in urban India, we demonstrate that new design elements – including simplified user interfaces, simplified task instructions, and language localization – are absolutely necessary to enable low-income workers to participate in and earn money using Mechanical Turk. We synthesize our findings into a set of design recommendations, as well as a realistic analysis of the potential for microtasking sites to deliver supplemental income to lower-income communities.