The microscope, as many other instruments, applies a linear filter to your measurement that cuts high frequency components that are “lost forever”. This apparently poses a fundamental limit to the information you can retrieve. SUPPOSe (from SUPerposition of Point Sources) is a deconvolution method that can “magically” recover part of the lost information from standard single shot images. The method is based on assuming that the object you want to reconstruct can be approximated by a superposition of virtual point sources of equal intensity. The problem of finding the intensity for each pixel is converted to that of locating the sources in a continuous space. The object does not need to be discrete, and the virtual sources can finally be blurred to represent a continuous object. The method will be presented and applications to super-resolution microscopy with fluorescent, multiphoton and scattering samples in 2D and 3D will be discussed.