The configuration of subvolcanic magma storage regions exercises a fundamental control on eruptive style and hazard. Such regions can be imaged remotely, using seismic, geodetic, or magnetotelluric methods, although these are far from routine and rarely unambiguous. The textures of erupted volcanic rocks, as quantified through crystal size distributions (CSD), provide space- and time-integrated information on subvolcanic plumbing systems, although these data cannot be used readily for reconstruction of key parameters such as conduit geometry or magma chamber depth. Here we develop a numerical approach to interpretation of CSD in products of steady eruptions, based on crystallization kinetics and hydrodynamic flow simulation, to image subvolcanic plumbing systems. The method requires knowledge of magma properties, crystal growth kinetics (measured experimentally), and discharge rate (measured observationally). The method is applicable to steady-state eruptive regimes. Distributions of pressure, temperature, crystal content, and conduit cross-section area with depth are obtained from a CSD from a sample erupted from Mount St. Helens volcano, USA. Values of average conduit diameter (∼30 m) and magma chamber depth (∼14 km below the summit) are in good agreement with independent estimates.