Incorporation of additional models to use would be beneficial for the end user and also the integration of HuggingFace models into the project, if required the client side or an extension to it could be added that can do the processing task on a local in-prem machine with open source text-to-image generation models as mentioned in the Title above.