Systems and methods relating to fragmenting content based on metadata are disclosed. In one embodiment, metadata is obtained from data that is accessible from an interface. A content descriptor, based on at least a portion of the metadata, may be utilized to determine fragment size. The content descriptor value may be derived from multiple forms of metadata, such as for example, genre, sub-genre, presence of live media, and combinations thereof. The fragments may be either virtual or physically discrete segments or files. In one aspect, the value may be the sole basis for determining the fragment lengths. In another aspect of the embodiment, the division of the content is conducted such that each resultant fragment comprises a single frame type.