The APPLgrid project provides a fast and flexible way to reproduce the results of full NLO calculations with any input parton distribution set in only a few milliseconds rather than the weeks normally required to gain adequate statistics. Written in C++ (although a fortran interface is included) it can be used for the calculation of any process where the hard subprocess weights from the convolution with the PDF are available from the calculation. The user can use existing grids simply to obtain the fast cross sections, as with fastNLO, but the complete project is pubically available should the user wish to generate the grids themselves for new cross sections. In this case, the user interfaces the grid code with NLO calculation, and after running the NLO calculation to achieve the required statistical precision once, the results of the calculation with any different parton distribution set can be calculated, typically in around 1 to 100ms, depending on the size of the grid. At present, examples exist for MCFM and nlojet++. Since the user code that needs to run to extract the weights and create the grid may require changes to the NLO calculation code or may be dependent on specific versions of the code, the specific versions of both MCFM and nlojet++ are included here An interface to fastNLO (version 1) grids is included. The code is under continuous development so please come back soon for more information.