Hi, your question is too broad, there are tons of models and it depends on which one you're referring. The only one that I know that could work with what you are describing is using a distilled SDXL model or lora with IP Adapters and img2img, but that's kind of old, nowadays you can use an edit model and for those models you need to ask the model what do you want, for example: change the style of image1 with the style of image2 which is a prompt, but if you're building an app, you can hide all of this and just make the user input two images.