Blue shifted absorption lines from highly ionised iron are seen in some high inclination X-ray binary systems, indicating the presence of an equatorial disc wind. This launch mechanism is under debate, but thermal driving should be ubiquitous. X-ray irradiation from the central source heats disc surface, forming a wind from the outer disc where the local escape velocity is lower than the sound speed. The mass loss rate from each part of the disc is determined by the luminosity and spectral shape of the central source. We use these together with an assumed density and velocity structure of the wind to predict the column density and ionisation state, then combine this with a Monte Carlo radiation transfer to predict the detailed shape of the absorption (and emission) line profiles. We test this on the persistent wind seen in the bright neutron star binary GX 13+1, with luminosity L/LEdd ~ 0.5. We approximately include the effect of radiation pressure because of high luminosity, and compute line features. We compare these to the highest resolution data, the Chandra third order grating spectra, which we show here for the first time. This is the first physical model for the wind in this system, and it succeeds in reproducing many of the features seen in the data, showing that the wind in GX13+1 is most likely a thermal-radiation driven wind. This approach, combined with better streamline structures derived from full radiation hydrodynamic simulations, will allow future calorimeter data to explore the detail wind structure.