In an audio-enhanced computer-assisted self-interview (audioCasi), respondents see questions on a computer screen while they hear them over headphones. Respondents with limited literacy skills potentially could benefit from the ability of audioCasi systems to present aural and graphical cues as well as text. The article describes the development of software for administering audioCasi surveys to both proficient readers and respondents with limited literacy skills. The initial version of the software was designed in accordance with principles drawn from the literature on computer-assisted interviewing and the human-computer interface. The software then underwent seven rounds of usability testing, a process in which people are observed using the software and then debriefed, with the goal of identifying deficiencies in the design of the software. People with a range of literacy skills participated in the usability tests. The software was revised following each round of testing. In the final version of the software, the screens were uncomplicated and the procedures involved in taking a survey were consistent and self-evident.