The arguments for converting sunlight and H2O to H2 to provide cleaner fuels and chemicals are very powerful. However, there is still no efficient means of direct solar energy conversion to H2 on a large scale despite a large research effort worldwide. This review describes strategies to develop robust devices which exploit the selectivity of a molecular catalyst but avoids the use of sacrificial electron donors by adsorbing them onto an electrode surface. By assembling the photocathodes with photoanodes, the electrons provided by water oxidation are used to reduce H+ to H2. By separating the functions of light absorption, charge transport and catalysis between the colloidal semiconductor and molecular components, the activity of each can be optimised. However, the complexity of the system requires advanced experimental techniques to evaluate the performance. Current understanding of the factors governing electron transfer across the interface between the semiconductor, dye and catalyst is described and future directions and challenges for this field are outlined.