Double deeply virtual Compton scattering (DDVCS) is a very precise tool for the nucleon tomography. Its
measurement requires high luminosity electron beams and precise dedicated detectors, since its amplitude
is quite small in the interesting kinematical domain where collinear QCD factorization allows the extraction
of quark and gluon generalized parton distributions (GPDs). We analyze the prospects for its study in the
JLab energy domain as well as in higher energy electron-ion colliders. Our results are very encouraging
for various observables both with an unpolarized and polarized lepton beam. Using various realistic
models for GPDs, we demonstrate that DDVCS measurements are indeed very sensitive to their behaviour.
Implementing our lowest order cross-section formulae in the EpIC Monte Carlo generator, we estimate the
expected number of interesting events.