A video clip storage and retrieval system whereby video clips, stored locally and/or at a more remote location, can be requested and retrieved by a user at the user's multimedia terminal. When the user requests a desired video clip, the request is processed by a primary index manager ("PIM") via a Local Search and Retrieval Unit ("SRU"). Before the message is communicated to the PIM, the local SRU checks its own storage to see whether the requested video clips are available locally. If some of the video clips are local, the local SRU still forwards the request to the PIM so that the PIM may determine specific video clip usage. The PIM determines the extended SRU where the audio-visual data is stored and passes this information to a Data Sequencing Interface ("DSI"). The DSI collects the video clips and downloads the clips to the user's terminal. The user may then view, copy, or print the video clip as desired. In a preferred embodiment, a distributed digital video clip delivery system, according to the invention, provides video clips stored at local and/or remote locations, which can be requested from the Internet and retrieved at the user's multimedia terminal. When the user requests a desired video clip shown on a Web page, the request is diverted to a primary index manager ("PIM"). The PIM attempts to locate the closest server containing the requested clip, from which the download is completed. The system further includes means for uploading and distributing clips to geographically diverse servers, dynamic load balancing, subscription management mechanisms, and protection means to discourage unauthorized duplication of video clips.