In this work, a framework is developed for the planning of an off-grid stand-alone PV system operating under a minimum Loss of Power Supply Probability (LPSP) requirement. In the first step, PV module and battery sizes are determined for a single house system(SHS) using an iterative numerical approach. A house may have a low (0.5kWh), medium (2kWh) or high (6kWh) daily average load demand. LPSP over one year is calculated for an array of values of PV module and battery sizes from which the optimal sizes are later selected based on the least levelized cost of energy (LCOE). Battery sizes are fixed and the PV module sizes are varied until the desired LPSP is reached. This helps to determine system sizes with a certain level of reliability (LPSP) while considering the high upfront costs associated with batteries. Component sizes are determined for LPSP limits of 10% and 5%. In the second step, the framework is applied to a multi-house system(MHS), whereby multiple houses are connected in various configurations, with the option of sharing excess energy. It is observed that connecting the houses with a similar load demand and component sizes results in an overall LPSP reduction. Net annual energy sharing profiles are observed for all the houses, revealing that more energy is received annually by the houses having larger battery sizes. Lastly, the LCOE of MHSs is compared against the single-house case, showing that when houses of similar load demand and component sizes are connected together, the LCOE is comparable to that of a SHS. This is explained by the reduction in LPSP obtained from such a configuration, thus providing room for a reduction in the component sizes. This work is useful for cost optimizations in offgrid regimes electrified through solar.