Combing through the news or any sort of media in terms of finding a topic of interest has been implemented in countries like China or America. In Pakistan, a relevant system, i.e., applicable both for Urdu and English, is being developed by a team at MDAR. This system has potential application in Regulatory bodies, media houses Intelligence agencies as well as general end users. The caption text appearing in videos is used as the primary index, while the audio content in these videos serves as the secondary index. The textual content appearing in videos in the form of News tickers, anchor names, alerts or any other information is localized and converted into text using a video Optical Character Recognition. The system considers Urdu and English text, but can be extended to other scripts as well. Examples of typical queries include Retrieve all videos where the keyword ‘صدر’ has appeared or Retrieve all videos where the keyword ‘پاکستان’ has been spoken, and others.