||IRIS Mg II h&k data have proved to be one of the best spectral rangeS to investigate the chromosphere and high photosphere. The inversion of these lines is the most suitable way to recover the thermodynamics in these regions. The STiC code is the best -and unique- available tool to extract this information. However, since it is computationally expensive, we have developed IRIS2 (IRIS-squared) that uses STiC inversions of representative profiles of IRIS Mg II h&k data to build a database of inverted profiles, their corresponding representative model atmospheres, and their response functions - which are eventually used to calculate uncertainties. Using this approach, IRIS2 speeds-up the inversion process by a factor of 10^5-10^6. In this talk, we present the current status of IRIS2, which is publicly available, and the python versions of IRIS2 and its deep learning variant, known as deepIRIS2. We will pay special attention to the first results obtained with IRIS2 inversions, the analysis of the uncertainties, and the upcoming new version of IRIS2. This new approach allows us to invert massive data in an efficient with accuracy that is comparable to that obtained with STiC or other traditional inversion codes. This method can be beneficial for projects like DKIST or EST.