In this workshop we will first define the basic glossary and main concepts of machine learning. We will then present Cellpose, a machine learning software for segmentation in open access with tow variants. One is an interactive online application which does not need neither a powerful computer or coding skills, the other a local installation which needs a bit of both. We will go through the requirement and installation of the second one, then start using its tools and functions with different kind of microscopy images. After that we will explore the notion of quality control to measure the accuracy of our segmentation and also how to use them in the most efficient and relevant way, including interfacing Cellpose with either ImageJ or Qupath. At the end of this workshop, users should be able to segment complex acquisitions within a couple of minutes, check their accuracy and relevance, and exploit them. Images will be supplied for the workshop, but participants can bring their own images if they choose so.