This paper outlines the creative and technical considerations behind AudioGuide1 , a program for differed-time concatenative synthesis written in Python. AudioGuide is a framework for experimentation with a ﬂexible concatenative algorithm. Two aspects of this algorithm are discussed in detail. First, strategies for ﬂexible feature mapping are presented which allow the user a degree of expressive freedom in shaping and transforming concatenated outputs. Second, a subtractive spectral algorithm is outlined which enables the selection of simultaneous corpus units. Simultaneous selection permits both vertically stratiﬁed units as well as horizontally overlapping units and, more generally, permits a more ﬂexible and comprehensive deployment of corpus resources. These two aspects, used in tandem, create a software framework capable of both sound synthesis and musical writing, enabling the user to move freely between sonic and score outputs.the sum of individual events.