This paper generalizes the strong seed-bank model introduced in arXiv:1411.4747 to allow for more general dormancy time distributions, such as a type of Pareto distribution. Inspired by the method of approximation using models with countably many seed-banks proposed by arXiv:2209.10086, we introduce the Wright-Fisher diffusion and coalescent with a continuum of seed-banks. To this end, we first formulate an infinite-dimensional stochastic differential equation, and prove that it has a unique strong solution, refereed to as the \textit{continuum seed-bank diffusion}, which is a kind of Markovian lift of a non-Markovian Volterra process. In order to circumvent the technical difficulty arising from the lack of local compactness, we replace the topology induced by the norm on the state space of the solution with the weak-$^{\star}$ topology, and show that the continuum seed-bank diffusion is also a strong Markov process in this weak-$^{\star}$ setting. Then, we construct a discrete-time Wright-Fisher type model with finitely many seed-banks, and demonstrate that the continuum seed-bank diffusion under the weak-$^{\star}$ topology is the scaling limit of the allele frequency processes in a suitable sequence of such models. Finally, we establish the duality relation between the continuum seed-bank diffusion and a continuous-time continuous-state Markov jump process. The latter is the block counting process of a partition-valued Markov jump process which is referred to as the \textit{continuum seed-bank coalescent}. We prove that this new coalescent process is exactly the scaling limit of the ancestral processes in the sequence of discrete-time Wright-Fisher type models we constructed before.