High-precision pulsar timing is central to a wide range of astrophysics and fundamental physics applications. When timing an ensemble of millisecond pulsars in different sky positions, known as a pulsar timing array (PTA), one can search for ultra-low-frequency gravitational waves (GWs) through the spatial correlations that spacetime deformations by passing GWs are predicted to induce on the pulses’ times-of-arrival (TOAs). A pulsar-timing model, requires the use of a solar-system ephemeris (SSE) to properly predict the position of the solar-system barycentre, the (quasi-)inertial frame where all TOAs are referred. Here, I discuss how while errors in SSEs can introduce correlations in the TOAs that may interfere with GW searches, one can make use of PTAs to study the solar system. I discuss work done within the context of the European Pulsar Timing Array and the International Pulsar Timing Array collaborations. These include new updates on the masses of planets from PTA data, first limits on masses of the most massive asteroids, and comparisons between SSEs from independent groups. Finally, I discuss a new approach in setting limits on the masses of unknown bodies in the solar system and calculate mass sensitivity curves for PTA data.