From sophisticated personal voice assistants like Siri or Alexa to simplistic keyword-based search bots, today, the label ``chatbot'' is used broadly for all kinds of systems that use natural language as input.
However, the systems summarized under this term are so diverse, that they often have very little in common with regard to technology, usage, and their theoretical background.
In order to make such systems more comparable, we propose a framework that classifies chatbots based on six categories,
which allow a meaningful comparison based on features which are relevant for developers, scientists, and users.
Ultimately, we hope to support the scientific discourse, as well as the development of chatbots, by providing an instrument to classify and analyze different groups of chatbot systems regarding their requirements, possible evaluation strategies, available toolsets, and other common features.