Supermassive stars (SMS) are massive hydrogen objects, slowly radiating their gravitational binding energy. Such hypothetical primordial objects may have been the seed of the massive black holes (BHs) observed at the centre of galaxies. Under the standard picture, these objects can be approximately described as n=3 polytropes, and they are expected to shine extremely close to their Eddington luminosity. Once however, one considers the porosity induced by instabilities near the Eddington limit, which give rise to super-Eddington states, the standard picture should be modified. We study the structure, evolution and mass loss of these objects. We find the following. First, the evolution of SMSs is hastened due to their increased energy release. They accelerate continuum driven winds. If there is no rotational stabilization, these winds are insufficient to "evaporate" the objects, such that they can collapse to form a supermassive BHs, however, they do prevent SMSs from emitting a copious amount of ionizing radiation. If the SMSs are rotationally stabilized, the winds "evaporate" the objects until a normal sub-Eddington star remains, having a mass of a few 100Msun.

We consider a scenario where supermassive black holes form through direct accumulation of gas at the centre of proto-galaxies. In the first stage, the accumulated gas forms a super-massive star whose core collapses when the nuclear fuel is exhausted, forming a black hole of $M_{\rm BH} \approx 100 M_{\sun}$. As the black hole starts accreting, it inflates the surrounding dense gas into an almost hydrostatic self-gravitating envelope, with at least 10-100 times the mass of the hole. We find that these "quasistars" suffer extremely high rates of mass loss through winds from their envelopes, in analogy to very massive stars such as eta-Carinae. Only for envelope masses greater than 2.8 \times 10^{5} (M_{\rm BH}/100 M_{\sun})^{9/11} is the envelope evaporation time-scale longer than the accretion time-scale of the black hole. This relation thus constitutes a "threshold growth line" above which quasistars can grow their internal black holes. Accretion rates can be 10 to 100 times the Eddington rate. The quasistars born in this "growth region" with 10^7-10^8 M_{\sun} can grow black holes with masses between 10^4$ to $10^5 M_{\sun}, before crossing the threshold growth line and dispersing their envelopes in less than $10^4$ yr. This scenario therefore predicts that massive black hole seeds can be found only in dark matter halos with total masses larger than about 10^9 M_{\sun}, which can provide sufficiently high accretion rates to form such massive quasistars.

We construct Super-Eddington Slim Disks models around both stellar and super-massive black holes by allowing the formation of a porous layer with a reduced effective opacity. We show that at high accretion rates, the inner part of the disks become radiation pressure dominated. However, unlike the standard scenario in which the disks become thick, super-Eddington disks remain slim. In addition, they accelerate a significant wind with a "thick disk" geometry. We show that above about 1.5 times the standard critical mass accretion rate (needed to release the Eddington luminosity), the net luminosity released is above Eddington. At above about 5 times the standard critical rate, the central BH accretes more than the Eddington accretion rate. Above about 20 m-dot_crit, the disk remains slim but the wind becomes spherical, and the present model breaks down.