which LLM are you using for those frontend usecases? chatgpt? and you ask in prompts for some framework such as tailwind?