“Building a toolbox for algorithmic, adaptive music” is a dissertation, describing the technical background of the computer program “Conductor”, used for generating adaptive music with algorithmic functionality from machine learning. Accepted approaches in algorithmic and adaptive composing are evaluated for their usability. Combinations of established and proprietary developed methods serve as a basis for an implementation. The result of the study is a prototype implementation. The program will use methods of machine learning to generalize input examples and generate a rule base. The rule base serves as a starting point for composing adaptive algorithmic music in real time.