Media clips on a hard drive or other storage are automatically organized in collections and metadata automatically assigned to facilitate subsequent searching for media clips. A Clip Organizer software application initially scans a user's hard drive to identify folders containing user files of sound, picture, clip art, video and other media clips. Predefined types of folders are omitted in the scan to avoid identifying media clips that do not belong to the user or are part of another application program. A partially "flattened" collection hierarchy is created to simplify browsing the media clips. Sufficient empty files are included to enable a user to readily locate the collections, and metadata comprising keywords derived from the path of each media clip are associated with each of the media clips to facilitate subsequent searching for the files. In addition, the collection hierarchy is automatically updated as changes are made.