A fundamental question in astrophysics is how stars get their mass. We know that low-mass stars form from the collapse of self-gravitating prestellar-cores. Since this collapse, young stellar objects (YSOs) acquire mass through the magnetospheric accretion process for up to 10Myr. According to this scenario, the material falls from the envelope through the circumstellar disk onto the central forming-star, following the magnetic field lines. Thanks to new facilities, it has been possible to observe spectroscopically the inner part of the circumstellar disks in the nerby star forming clouds. Therefore, accurate estimates of the mass accretion rate (Macc) and stellar parameters in different stages (early, i.e. ClassI, and more evolved, i.e. Classical T-Tauri stars) of the star formation process have been provided for single stars and binaries. However, if we integrate Macc provided from the observations for the estimated timescales of YSOs, we found smaller masses than we measure. This means that the majority of the mass is set during the first stage of the highly embedded protostellar phase (Class0), where planets start to form, or the accretion process proceeds in a non-steady framework. While we still know much less on accretion on Class0, the non-steady accretion is proven by the eruptive YSOs, as FUors and EXors, which experience extremely strong bursts on short and long timescales. I will review recent results about accretion, focusing on open questions on early stages, as how the forming-star mass is related to the disk and envelope mass, and the relation between models and observations.