Creating new multimedia content is normally a task out of the reach of automatic tools. The exception is the creation of video summaries, using an existing video as source. In this paper, a tool for creating new video content to fit a determinate purpose is introduced. It uses a Mpeg-7 database of video descriptors and annotations to select the segments that best fit that purpose. The required descriptors can be automatically extracted, but additional manual annotations can be included to pre-filter the database using application domain knowledge. The tool is able to use segments from several videos and automates all the creation process, following the parameters defined by a supervisor in a program specification document. This document provides enough flexibility to customize the new program to whatever purpose is intended, with parameters to set low level features (color, motion,...) as well as production related parameters.