Recursive sequences of laws of random variables (and random vectors) are considered where an independence assumption which is usually made within the setting of the contraction method is dropped. This restricts the study to sequences which after normalization lead to asymptotic normality. We provide a general univariate central limit theorem which can directly be applied to problems from the analysis of algorithms and random recursive structures without further knowledge of the contraction method. Also multivariate central limit theorems are shown and bounds on rates of convergence are provided. Examples include some previously shown central limit analogues as well as new applications on Fibonacci matchings.