The inherent irreversibility of quantum dynamics for open systems poses a significant barrier to the inversion of unknown quantum processes. To tackle this challenge, we propose the framework of virtual combs that exploit the unknown process iteratively with additional classical post-processing to simulate the process inverse. Our research establishes a path to achieving the exact inverse of unknown channels with certain conditions, accompanied by a no-go theorem that underscores the intrinsic limitations imposed by quantum mechanics on such tasks. Notably, we demonstrate that an n-slot virtual comb can exactly reverse a depolarizing channel with one unknown noise parameter out of n+1 potential candidates, and a 1-slot virtual comb can exactly reverse an arbitrary pair of quantum channels. We further explore the approximate inverse of an unknown channel within a given channel set. For any unknown depolarizing channels within a specified noise region, we unveil a worst-case error decay of O(n^(-1)) of reversing the channel via virtual combs. Moreover, we show that virtual combs with constant slots can be applied to universally reverse unitary operations and investigate the trade-off between the slot number and the sampling overhead.