Space interferometers could, in principle, exploit the relatively stable space environment and ease of baseline reconfiguration to collect measurements beyond the limitations of ground-based interferometers. In particular, a two-element interferometer could provide excellent uv-plane coverage over a few tens of low Earth orbits. One of the challenges for free-flying interferometers is controlling the optical path distance with subwavelength accuracies despite the collectors flying up to hundreds of meters apart. We consider two approaches: an artificial in-orbit laser guide star (LGS) that provides a phase reference for the space interferometer and fringe tracking on the science target itself. The two approaches (LGS versus no LGS) would require different image processing techniques. In this work, we explore image processing with LGS phase residuals due to global positioning system (GPS) uncertainties. We use GPS uncertainties from the Gravity Recovery and Climate Experiment Follow-On mission to simulate image retrieval with a 300-m baseline laser-guided space interferometer. This is done by fitting the slowly varying phase errors of complex visibility measurements. We also consider a 40-m baseline interferometer with visibility(-modulus)-only measurements. In this case, we simulate the bias in visibility due to fringe tracking in the presence of parasitic forces acting on the spacecraft. We then use a modified version of the hybrid input–output phase retrieval algorithm for image reconstruction. We conclude that under our optimistic assumptions, both approaches could enable general imaging of a few large stars even with CubeSats, although an LGS would significantly improve the best resolution obtainable.