This paper reports on the potential of Oral Elicited Imitation (OEI) as a format for output practice, building on an analysis of picture-matching and spoken data collected from 36 university-level learners of German as a second language (L2) in a web-based assessment task inspired by Input Processing (VanPatten, 2004). The design and development of OEI for output practice faces two key challenges: learners must be engaged in meaningful language processing rather than in mere repetition of oral stimuli, and the task must eventually provide individualized and qualitative corrective feedback that helps learners to notice gaps between their interlanguage and the target language. Results show that learners attended to meaning and that a commercially available speech recognition tool was able to transcribe learner speech remarkably well. [For the complete volume, see ED578177.]