Modeling of extreme weather data typically involves thresholding, or retaining only the most extreme observations, in order that the tail may "speak" and not be overwhelmed by the bulk of the data. We propose a transformation-based method that smoothly transitions from a flexible, semi-parametric estimation of the bulk to a parametric estimation of the tails without thresholding. In the limit, this method has desirable theoretical tail-matching properties. We carry out theoretical and empirical investigation to explore its practical usage at pre-limit extremes (e.g. estimating the hundred-year flood instead of the billion-year flood), comparing its performance to methods that involve thresholding.