There are models available to run text to image generation and image to image generation with text prompt guidance but I dont think they can create ghibli art. You may need to finetune those models to get the results. A reference repo
easy,
and and there are probably dozens loras along with ideally suited checkpoints of them out there,
and you can run them on ai-img gen software like comfyui/sd/a1111/fooocus and similar
