Video-based concordancer can be used to provide a scenario that engage context information of studying novel words. Usually, a fixed number of contextual sentences are retrieved accompanying the keywords. However, there may be a lack of complete context for learners to comprehend the keywords in the videos. Few studies have discussed about how the videos are presented to assist learners to use the keywords appropriately, and lead learners to find relevant knowledge effectively. In this paper, a keyword-inscene video concordance (KWIS), which recognizes the scenes in the videos and provides the scene-based clips, is proposed. Each video clip is tagged actual scene type information. Learners are able to query the KWIS system with keywords, phrases, or natural language sentences, and watch relevant scenario clips to understand where the conversation can be carried on. A pilot study was conducted to evaluate the proposed system. The result shows that there is a positive effect on students' comprehension of English phrases while using the system.