In a series of recent papers we have developed what we call the DEKI account of scientific representation, according to which models represent their targets via keys. These keys provide a systematic way to move from model-features to features to be imputed to their targets. We show how keys allow for accurate representation in the presence of idealisation, and further illustrate how investigating them provides novel ways to approach certain currently debated questions in the philosophy of science. To add specificity, we offer a detailed analysis of a kind of key that that is crucial in many parts of physics, namely what we call limit keys. These keys exploit the fact that the features exemplified by these models are limits of the features of the target.