Mi Dai, Jeremy Kubica, Konstantin Malanchev, Alex I. Malz, Olivia Lynn, Andrew Connolly, Rachel Mandelbaum, W. M. Wood-Vasey

We present LightCurveLynx, a flexible and extensible software framework for end-to-end forward modeling time-domain light curves. Given the growing need for realistic simulations in the time-domain astronomy community, LightCurveLynx is designed to support a wide range of applications, including the development and validation of analysis pipelines, the optimization of survey strategies, and simulation-based inference studies. Realistic simulations can be generated from real survey metadata, forecasted survey plans, or user-defined mock survey strategies. We demonstrate the functionality of LightCurveLynx by generating a realistic simulation of Type Ia supernovae that is representative of the ZTF SN Ia Data Release 2 dataset and perform extensive comparisons between the simulated and observed samples to validate the software. The simulation shows excellent agreement with the data in parameter distributions (with the Kullback-Leibler divergence values around 0.01-0.02) and in noise properties. The Hubble diagram generated from the simulation also indicates that the sample is complete up to redshift 0.06, which is consistent with previous studies. Our results confirm that LightCurveLynx is robust, accurate, and ready for community use and contribution.